freebsd-skq/sys/kern/vfs_extattr.c

/*
 * Copyright (c) 1989, 1993
 *	The Regents of the University of California.  All rights reserved.
 * (c) UNIX System Laboratories, Inc.
 * All or some portions of this file are derived from material licensed
 * to the University of California by American Telephone and Telegraph
 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
 * the permission of UNIX System Laboratories, Inc.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	@(#)vfs_syscalls.c	8.13 (Berkeley) 4/15/94
 * $FreeBSD$
 */

/* For 4.3 integer FS ID compatibility */
#include "opt_compat.h"
#include "opt_ffs.h"

#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bio.h>
#include <sys/buf.h>
#include <sys/sysent.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/mutex.h>
#include <sys/sysproto.h>
#include <sys/namei.h>
#include <sys/filedesc.h>
#include <sys/kernel.h>
#include <sys/fcntl.h>
#include <sys/file.h>
#include <sys/linker.h>
#include <sys/stat.h>
#include <sys/sx.h>
#include <sys/unistd.h>
#include <sys/vnode.h>
#include <sys/proc.h>
#include <sys/dirent.h>
#include <sys/extattr.h>
#include <sys/jail.h>
#include <sys/sysctl.h>

#include <machine/limits.h>
#include <machine/stdarg.h>

#include <vm/vm.h>
#include <vm/vm_object.h>
#include <vm/vm_page.h>
#include <vm/uma.h>

static int change_dir(struct nameidata *ndp, struct thread *td);
static void checkdirs(struct vnode *olddp, struct vnode *newdp);
static int chroot_refuse_vdir_fds(struct filedesc *fdp);
static int getutimes(const struct timeval *, struct timespec *);
static int setfown(struct thread *td, struct vnode *, uid_t, gid_t);
static int setfmode(struct thread *td, struct vnode *, int);
static int setfflags(struct thread *td, struct vnode *, int);
static int setutimes(struct thread *td, struct vnode *,
    const struct timespec *, int);
static int vn_access(struct vnode *vp, int user_flags, struct ucred *cred,
    struct thread *td);
static void vfs_freeopts(struct vfsoptlist *opt);
static int vfs_nmount(struct thread *td, int, struct uio *);

static int	usermount = 0;	/* if 1, non-root can mount fs. */

int (*union_dircheckp)(struct thread *td, struct vnode **, struct file *);

SYSCTL_INT(_vfs, OID_AUTO, usermount, CTLFLAG_RW, &usermount, 0, "");

/*
 * Virtual File System System Calls
 */

#ifndef _SYS_SYSPROTO_H_
struct nmount_args {
	struct iovec    *iovp;
	unsigned int    iovcnt;
	int             flags;
};
#endif
/* ARGSUSED */
int
nmount(td, uap)
	struct thread *td;
	struct nmount_args /* {
		syscallarg(struct iovec *) iovp;
		syscallarg(unsigned int) iovcnt;
		syscallarg(int) flags;
	} */ *uap;
{
	struct uio auio;
	struct iovec *iov, *needfree;
	struct iovec aiov[UIO_SMALLIOV];
	long error, i;
	u_int iovlen, iovcnt;

	iovcnt = SCARG(uap, iovcnt);
	iovlen = iovcnt * sizeof (struct iovec);
	/*
	 * Check that we have an even number of iovec's
	 * and that we have at least two options.
	 */
	if ((iovcnt & 1) || (iovcnt < 4) || (iovcnt > UIO_MAXIOV))
		return (EINVAL);

	if (iovcnt > UIO_SMALLIOV) {
		MALLOC(iov, struct iovec *, iovlen, M_IOV, M_WAITOK);
		needfree = iov;
	} else {
		iov = aiov;
		needfree = NULL;
	}
	auio.uio_iov = iov;
	auio.uio_iovcnt = iovcnt;
	auio.uio_rw = UIO_WRITE;
	auio.uio_segflg = UIO_USERSPACE;
	auio.uio_td = td;
	auio.uio_offset = 0;
	auio.uio_resid = 0;
	if ((error = copyin((caddr_t)uap->iovp, (caddr_t)iov, iovlen)))
		goto finish;
	for (i = 0; i < iovcnt; i++) {
		if (iov->iov_len > INT_MAX - auio.uio_resid) {
			error = EINVAL;
			goto finish;
		}
		auio.uio_resid += iov->iov_len;
		iov++;
	}
	error = vfs_nmount(td, SCARG(uap, flags), &auio);
finish:
	if (needfree != NULL)
		free(needfree, M_TEMP);
	return (error);
}

/*
 * Release all resources related to the
 * mount options.
 */
static void
vfs_freeopts(struct vfsoptlist *opt)
{
	free(opt->opt, M_MOUNT);
	free(opt->optbuf, M_MOUNT);
	free(opt, M_MOUNT);
}

int
kernel_mount(iovp, iovcnt, flags)
	struct iovec *iovp;
	unsigned int iovcnt;
	int flags;
{
	struct uio auio;
	struct iovec *iov;
	int error, i;

	/*
	 * Check that we have an even number of iovec's
	 * and that we have at least two options.
	 */
	if ((iovcnt & 1) || (iovcnt < 4))
		return (EINVAL);

	auio.uio_iov = iovp;
	auio.uio_iovcnt = iovcnt;
	auio.uio_rw = UIO_WRITE;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_offset = 0;
	auio.uio_td = NULL;
	auio.uio_resid = 0;
	iov = iovp;
	for (i = 0; i < iovcnt; i++) {
		if (iov->iov_len > INT_MAX - auio.uio_resid) {
			return (EINVAL);
		}
		auio.uio_resid += iov->iov_len;
		iov++;
	}

	error = vfs_nmount(curthread, flags, &auio);
	return (error);
}

int
kernel_vmount(int flags, ...)
{
	struct iovec *iovp, *iov;
	struct uio auio;
	va_list ap;
	unsigned int iovcnt, iovlen, len, optcnt;
	const char *opt;
	char *sep, *buf, *pos;
	int error, i;

	len = 0;
	va_start(ap, flags);
	for (optcnt = 0; (opt = va_arg(ap, const char *)) != NULL; optcnt++)
		len += strlen(opt) + 1;
	va_end(ap);

	if (optcnt < 2)
		return (EINVAL);

	iovcnt = optcnt << 1;
	iovlen = iovcnt * sizeof (struct iovec);
	MALLOC(iovp, struct iovec *, iovlen, M_MOUNT, M_WAITOK);
	MALLOC(buf, char *, len, M_MOUNT, M_WAITOK);
	pos = buf;
	va_start(ap, flags);
	for (i = 0; i < optcnt; i++) {
		opt = va_arg(ap, const char *);
		strcpy(pos, opt);
		sep = index(pos, '=');
		if (sep == NULL) {
			FREE(iovp, M_MOUNT);
			FREE(buf, M_MOUNT);
			return (EINVAL);
		}
		*sep = '\0';
		iov = iovp + i * 2;
		iov->iov_base = pos;
		iov->iov_len = sep - pos + 1;
		pos = sep + 1;
		iov++;
		iov->iov_base = pos;
		iovlen = strlen(pos) + 1;
		iov->iov_len = iovlen;
		pos += iovlen;
	}
	va_end(ap);

	auio.uio_iov = iovp;
	auio.uio_iovcnt = iovcnt;
	auio.uio_rw = UIO_WRITE;
	auio.uio_segflg = UIO_SYSSPACE;
	auio.uio_offset = 0;
	auio.uio_td = NULL;
	auio.uio_resid = len;

	error = vfs_nmount(curthread, flags, &auio);
	FREE(iovp, M_MOUNT);
	FREE(buf, M_MOUNT);
	return (error);
}

/*
 * vfs_nmount(): actually attempt a filesystem mount.
 */
static int
vfs_nmount(td, fsflags, fsoptions)
	struct thread *td;
	int fsflags;		/* Flags common to all filesystems */
	struct uio *fsoptions;	/* Options local to the filesystem */
{
	struct vnode *vp;
	struct mount *mp;
	struct vfsconf *vfsp;
	struct iovec *cur;
	struct vfsoptlist *optlist;
	struct vfsopt *opt;
	char *buf, *fstype, *fspath;
	int error, flag = 0, flag2 = 0, i, len, optcnt;
	int offset, iovcnt, fstypelen, fspathlen;
	struct vattr va;
	struct nameidata nd;

	/*
	 * Allocate memory to hold the vfsopt structures.
	 */
	iovcnt = fsoptions->uio_iovcnt;
	optcnt = iovcnt >> 1;
	opt = malloc(sizeof (struct vfsopt) * optcnt,
	    M_MOUNT, M_WAITOK | M_ZERO);

	/*
	 * Count the size of the buffer for options,
	 * allocate it, and fill in the vfsopt structures.
	 */
	cur = fsoptions->uio_iov;
	len = fsoptions->uio_resid;
	buf = malloc(len, M_TEMP, M_WAITOK | M_ZERO);

	optlist = malloc(sizeof (struct vfsoptlist), M_MOUNT, M_WAITOK);
	optlist->opt = opt;
	optlist->optbuf = buf;
	optlist->optcnt = optcnt;

	offset = i = 0;
	cur = fsoptions->uio_iov;
	while (i < optcnt) {
		opt[i].name = buf + offset;
		/* Ensure the name of an option is a string */
		if (opt[i].name[cur->iov_len - 1] != '\0') {
			error = EINVAL;
			goto bad;
		}
		offset += cur->iov_len;
		cur++;
		opt[i].len = cur->iov_len;
		/*
		 * Prevent consumers from trying to
		 * read the value of a 0 length option
		 * by setting it to NULL.
		 */
		if (opt[i].len == 0)
			opt[i].value = NULL;
		else
			opt[i].value = buf + offset;
		offset += cur->iov_len;
		cur++; i++;
	}

	if ((error = uiomove(buf, len, fsoptions)) != 0)
		goto bad;

	/*
	 * We need these two options before the others,
	 * and they are mandatory for any filesystem.
	 * Ensure they are NULL terminated as well.
	 */
	fstypelen = 0;
	error = vfs_getopt(optlist, "fstype", (void **)&fstype, &fstypelen);
	if ((error != 0) || (fstype[fstypelen - 1] != '\0')) {
		error = EINVAL;
		goto bad;
	}
	fspathlen = 0;
	error = vfs_getopt(optlist, "fspath", (void **)&fspath, &fspathlen);
	if ((error != 0) || (fspath[fspathlen - 1] != '\0')) {
		error = EINVAL;
		goto bad;
	}

	/*
	 * Be ultra-paranoid about making sure the type and fspath
	 * variables will fit in our mp buffers, including the
	 * terminating NUL.
	 */
	if ((fstypelen >= MFSNAMELEN - 1) ||
	    (fspathlen >= MNAMELEN - 1)) {
		error = ENAMETOOLONG;
		goto bad;
	}

	if (usermount == 0) {
	       	error = suser_td(td);
		if (error)
			goto bad;
	}
	/*
	 * Do not allow NFS export by non-root users.
	 */
	if (fsflags & MNT_EXPORTED) {
		error = suser_td(td);
		if (error)
			goto bad;
	}
	/*
	 * Silently enforce MNT_NOSUID and MNT_NODEV for non-root users
	 */
	if (suser_xxx(td->td_proc->p_ucred, 0, 0)) 
		fsflags |= MNT_NOSUID | MNT_NODEV;
	/*
	 * Get vnode to be covered
	 */
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, fspath, td);
	if ((error = namei(&nd)) != 0)
		goto bad;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;
	if (fsflags & MNT_UPDATE) {
		if ((vp->v_flag & VROOT) == 0) {
			vput(vp);
			error = EINVAL;
			goto bad;
		}
		mp = vp->v_mount;
		flag = mp->mnt_flag;
		flag2 = mp->mnt_kern_flag;
		/*
		 * We only allow the filesystem to be reloaded if it
		 * is currently mounted read-only.
		 */
		if ((fsflags & MNT_RELOAD) &&
		    ((mp->mnt_flag & MNT_RDONLY) == 0)) {
			vput(vp);
			error = EOPNOTSUPP;	/* Needs translation */
			goto bad;
		}
		/*
		 * Only root, or the user that did the original mount is
		 * permitted to update it.
		 */
		if (mp->mnt_stat.f_owner != td->td_proc->p_ucred->cr_uid) {
			error = suser_td(td);
			if (error) {
				vput(vp);
				goto bad;
			}
		}
		if (vfs_busy(mp, LK_NOWAIT, 0, td)) {
			vput(vp);
			error = EBUSY;
			goto bad;
		}
		mtx_lock(&vp->v_interlock);
		if ((vp->v_flag & VMOUNT) != 0 ||
		    vp->v_mountedhere != NULL) {
			mtx_unlock(&vp->v_interlock);
			vfs_unbusy(mp, td);
			vput(vp);
			error = EBUSY;
			goto bad;
		}
		vp->v_flag |= VMOUNT;
		mtx_unlock(&vp->v_interlock);
		mp->mnt_flag |= fsflags &
		    (MNT_RELOAD | MNT_FORCE | MNT_UPDATE | MNT_SNAPSHOT);
		VOP_UNLOCK(vp, 0, td);
		mp->mnt_optnew = optlist;
		goto update;
	}
	/*
	 * If the user is not root, ensure that they own the directory
	 * onto which we are attempting to mount.
	 */
	error = VOP_GETATTR(vp, &va, td->td_proc->p_ucred, td);
	if (error) {
		vput(vp);
		goto bad;
	}
	if (va.va_uid != td->td_proc->p_ucred->cr_uid) {
		error = suser_td(td);
		if (error) {
			vput(vp);
			goto bad;
		}
	}
	if ((error = vinvalbuf(vp, V_SAVE, td->td_proc->p_ucred, td, 0, 0))
	    != 0) {
		vput(vp);
		goto bad;
	}
	if (vp->v_type != VDIR) {
		vput(vp);
		error = ENOTDIR;
		goto bad;
	}
	for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next)
		if (!strcmp(vfsp->vfc_name, fstype))
			break;
	if (vfsp == NULL) {
		linker_file_t lf;

		/* Only load modules for root (very important!) */
		error = suser_td(td);
		if (error) {
			vput(vp);
			goto bad;
		}
		error = securelevel_gt(td->td_ucred, 0);
		if (error != 0) {
			vput(vp);
			goto bad;
		}
		error = linker_load_file(fstype, &lf);
		if (error || lf == NULL) {
			vput(vp);
			if (lf == NULL)
				error = ENODEV;
			goto bad;
		}
		lf->userrefs++;
		/* lookup again, see if the VFS was loaded */
		for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next)
			if (!strcmp(vfsp->vfc_name, fstype))
				break;
		if (vfsp == NULL) {
			lf->userrefs--;
			linker_file_unload(lf);
			vput(vp);
			error = ENODEV;
			goto bad;
		}
	}

	mtx_lock(&vp->v_interlock);
	if ((vp->v_flag & VMOUNT) != 0 ||
	    vp->v_mountedhere != NULL) {
		mtx_unlock(&vp->v_interlock);
		vput(vp);
		error = EBUSY;
		goto bad;
	}
	vp->v_flag |= VMOUNT;
	mtx_unlock(&vp->v_interlock);

	/*
	 * Allocate and initialize the filesystem.
	 */
	mp = malloc(sizeof(struct mount), M_MOUNT, M_WAITOK | M_ZERO);
	TAILQ_INIT(&mp->mnt_nvnodelist);
	TAILQ_INIT(&mp->mnt_reservedvnlist);
	lockinit(&mp->mnt_lock, PVFS, "vfslock", 0, LK_NOPAUSE);
	vfs_busy(mp, LK_NOWAIT, 0, td);
	mp->mnt_op = vfsp->vfc_vfsops;
	mp->mnt_vfc = vfsp;
	vfsp->vfc_refcount++;
	mp->mnt_stat.f_type = vfsp->vfc_typenum;
	mp->mnt_flag |= vfsp->vfc_flags & MNT_VISFLAGMASK;
	strncpy(mp->mnt_stat.f_fstypename, fstype, MFSNAMELEN);
	mp->mnt_stat.f_fstypename[MFSNAMELEN - 1] = '\0';
	mp->mnt_vnodecovered = vp;
	mp->mnt_stat.f_owner = td->td_proc->p_ucred->cr_uid;
	strncpy(mp->mnt_stat.f_mntonname, fspath, MNAMELEN);
	mp->mnt_stat.f_mntonname[MNAMELEN - 1] = '\0';
	mp->mnt_iosize_max = DFLTPHYS;
	VOP_UNLOCK(vp, 0, td);

	mp->mnt_opt = optlist;
update:
	/*
	 * Check if the fs implements the new VFS_NMOUNT()
	 * function, since the new system call was used.
	 */
	if (mp->mnt_op->vfs_mount != NULL) {
		printf("%s doesn't support the new mount syscall\n",
		    mp->mnt_vfc->vfc_name);
		vput(vp);
		error = EOPNOTSUPP;
		goto bad;
	}

	/*
	 * Set the mount level flags.
	 */
	if (fsflags & MNT_RDONLY)
		mp->mnt_flag |= MNT_RDONLY;
	else if (mp->mnt_flag & MNT_RDONLY)
		mp->mnt_kern_flag |= MNTK_WANTRDWR;
	mp->mnt_flag &=~ (MNT_NOSUID | MNT_NOEXEC | MNT_NODEV |
	    MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC | MNT_NOATIME |
	    MNT_NOSYMFOLLOW | MNT_IGNORE |
	    MNT_NOCLUSTERR | MNT_NOCLUSTERW | MNT_SUIDDIR);
	mp->mnt_flag |= fsflags & (MNT_NOSUID | MNT_NOEXEC |
	    MNT_NODEV | MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC | MNT_FORCE |
	    MNT_NOSYMFOLLOW | MNT_IGNORE |
	    MNT_NOATIME | MNT_NOCLUSTERR | MNT_NOCLUSTERW | MNT_SUIDDIR);
	/*
	 * Mount the filesystem.
	 * XXX The final recipients of VFS_MOUNT just overwrite the ndp they
	 * get.  No freeing of cn_pnbuf.
	 */
	error = VFS_NMOUNT(mp, &nd, td);
	if (mp->mnt_flag & MNT_UPDATE) {
		if (mp->mnt_kern_flag & MNTK_WANTRDWR)
			mp->mnt_flag &= ~MNT_RDONLY;
		mp->mnt_flag &=~
		    (MNT_UPDATE | MNT_RELOAD | MNT_FORCE | MNT_SNAPSHOT);
		mp->mnt_kern_flag &=~ MNTK_WANTRDWR;
		if (error) {
			mp->mnt_flag = flag;
			mp->mnt_kern_flag = flag2;
			vfs_freeopts(mp->mnt_optnew);
		} else {
			vfs_freeopts(mp->mnt_opt);
			mp->mnt_opt = mp->mnt_optnew;
		}
		if ((mp->mnt_flag & MNT_RDONLY) == 0) {
			if (mp->mnt_syncer == NULL)
				error = vfs_allocate_syncvnode(mp);
		} else {
			if (mp->mnt_syncer != NULL)
				vrele(mp->mnt_syncer);
			mp->mnt_syncer = NULL;
		}
		vfs_unbusy(mp, td);
		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		mtx_unlock(&vp->v_interlock);
		vrele(vp);
		return (error);
	}
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	/*
	 * Put the new filesystem on the mount list after root.
	 */
	cache_purge(vp);
	if (!error) {
		struct vnode *newdp;

		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		vp->v_mountedhere = mp;
		mtx_unlock(&vp->v_interlock);
		mtx_lock(&mountlist_mtx);
		TAILQ_INSERT_TAIL(&mountlist, mp, mnt_list);
		mtx_unlock(&mountlist_mtx);
		if (VFS_ROOT(mp, &newdp))
			panic("mount: lost mount");
		checkdirs(vp, newdp);
		vput(newdp);
		VOP_UNLOCK(vp, 0, td);
		if ((mp->mnt_flag & MNT_RDONLY) == 0)
			error = vfs_allocate_syncvnode(mp);
		vfs_unbusy(mp, td);
		if ((error = VFS_START(mp, 0, td)) != 0) {
			vrele(vp);
			goto bad;
		}
	} else {
		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		mtx_unlock(&vp->v_interlock);
		mp->mnt_vfc->vfc_refcount--;
		vfs_unbusy(mp, td);
		free((caddr_t)mp, M_MOUNT);
		vput(vp);
		goto bad;
	}
	return (0);
bad:
	vfs_freeopts(optlist);
	return (error);
}

/*
 * Old Mount API
 */
#ifndef _SYS_SYSPROTO_H_
struct mount_args {
	char	*type;
	char	*path;
	int	flags;
	caddr_t	data;
};
#endif
/* ARGSUSED */
int
mount(td, uap)
	struct thread *td;
	struct mount_args /* {
		syscallarg(char *) type;
		syscallarg(char *) path;
		syscallarg(int) flags;
		syscallarg(caddr_t) data;
	} */ *uap;
{
	char *fstype;
	char *fspath;
	int error;

	fstype = malloc(MFSNAMELEN, M_TEMP, M_WAITOK | M_ZERO);
	fspath = malloc(MNAMELEN, M_TEMP, M_WAITOK | M_ZERO);

	/*
	 * vfs_mount() actually takes a kernel string for `type' and
	 * `path' now, so extract them.
	 */
	error = copyinstr(SCARG(uap, type), fstype, MFSNAMELEN, NULL);
	if (error)
		goto finish;
	error = copyinstr(SCARG(uap, path), fspath, MNAMELEN, NULL);
	if (error)
		goto finish;
	error = vfs_mount(td, fstype, fspath, SCARG(uap, flags),
	    SCARG(uap, data));
finish:
	free(fstype, M_TEMP);
	free(fspath, M_TEMP);
	return (error);
}

/*
 * vfs_mount(): actually attempt a filesystem mount.
 *
 * This routine is designed to be a "generic" entry point for routines
 * that wish to mount a filesystem. All parameters except `fsdata' are
 * pointers into kernel space. `fsdata' is currently still a pointer
 * into userspace.
 */
int
vfs_mount(td, fstype, fspath, fsflags, fsdata)
	struct thread *td;
	const char *fstype;
	char *fspath;
	int fsflags;
	void *fsdata;
{
	struct vnode *vp;
	struct mount *mp;
	struct vfsconf *vfsp;
	int error, flag = 0, flag2 = 0;
	struct vattr va;
	struct nameidata nd;

	/*
	 * Be ultra-paranoid about making sure the type and fspath
	 * variables will fit in our mp buffers, including the
	 * terminating NUL.
	 */
	if ((strlen(fstype) >= MFSNAMELEN - 1) ||
	    (strlen(fspath) >= MNAMELEN - 1))
		return (ENAMETOOLONG);

	if (usermount == 0) {
		error = suser_td(td);
		if (error)
			return (error);
	}
	/*
	 * Do not allow NFS export by non-root users.
	 */
	if (fsflags & MNT_EXPORTED) {
		error = suser_td(td);
		if (error)
			return (error);
	}
	/*
	 * Silently enforce MNT_NOSUID and MNT_NODEV for non-root users
	 */
	if (suser_xxx(td->td_ucred, 0, 0)) 
		fsflags |= MNT_NOSUID | MNT_NODEV;
	/*
	 * Get vnode to be covered
	 */
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, fspath, td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;
	if (fsflags & MNT_UPDATE) {
		if ((vp->v_flag & VROOT) == 0) {
			vput(vp);
			return (EINVAL);
		}
		mp = vp->v_mount;
		flag = mp->mnt_flag;
		flag2 = mp->mnt_kern_flag;
		/*
		 * We only allow the filesystem to be reloaded if it
		 * is currently mounted read-only.
		 */
		if ((fsflags & MNT_RELOAD) &&
		    ((mp->mnt_flag & MNT_RDONLY) == 0)) {
			vput(vp);
			return (EOPNOTSUPP);	/* Needs translation */
		}
		/*
		 * Only root, or the user that did the original mount is
		 * permitted to update it.
		 */
		if (mp->mnt_stat.f_owner != td->td_ucred->cr_uid) {
			error = suser_td(td);
			if (error) {
				vput(vp);
				return (error);
			}
		}
		if (vfs_busy(mp, LK_NOWAIT, 0, td)) {
			vput(vp);
			return (EBUSY);
		}
		mtx_lock(&vp->v_interlock);
		if ((vp->v_flag & VMOUNT) != 0 ||
		    vp->v_mountedhere != NULL) {
			mtx_unlock(&vp->v_interlock);
			vfs_unbusy(mp, td);
			vput(vp);
			return (EBUSY);
		}
		vp->v_flag |= VMOUNT;
		mtx_unlock(&vp->v_interlock);
		mp->mnt_flag |= fsflags &
		    (MNT_RELOAD | MNT_FORCE | MNT_UPDATE | MNT_SNAPSHOT);
		VOP_UNLOCK(vp, 0, td);
		goto update;
	}
	/*
	 * If the user is not root, ensure that they own the directory
	 * onto which we are attempting to mount.
	 */
	error = VOP_GETATTR(vp, &va, td->td_ucred, td);
	if (error) {
		vput(vp);
		return (error);
	}
	if (va.va_uid != td->td_ucred->cr_uid) {
		error = suser_td(td);
		if (error) {
			vput(vp);
			return (error);
		}
	}
	if ((error = vinvalbuf(vp, V_SAVE, td->td_ucred, td, 0, 0))
	    != 0) {
		vput(vp);
		return (error);
	}
	if (vp->v_type != VDIR) {
		vput(vp);
		return (ENOTDIR);
	}
	for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next)
		if (!strcmp(vfsp->vfc_name, fstype))
			break;
	if (vfsp == NULL) {
		linker_file_t lf;

		/* Only load modules for root (very important!) */
		error = suser_td(td);
		if (error) {
			vput(vp);
			return error;
		}
		error = securelevel_gt(td->td_ucred, 0);
		if (error != 0) {
			vput(vp);
			return (EPERM);
		}
		error = linker_load_file(fstype, &lf);
		if (error || lf == NULL) {
			vput(vp);
			if (lf == NULL)
				error = ENODEV;
			return error;
		}
		lf->userrefs++;
		/* lookup again, see if the VFS was loaded */
		for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next)
			if (!strcmp(vfsp->vfc_name, fstype))
				break;
		if (vfsp == NULL) {
			lf->userrefs--;
			linker_file_unload(lf);
			vput(vp);
			return (ENODEV);
		}
	}
	mtx_lock(&vp->v_interlock);
	if ((vp->v_flag & VMOUNT) != 0 ||
	    vp->v_mountedhere != NULL) {
		mtx_unlock(&vp->v_interlock);
		vput(vp);
		return (EBUSY);
	}
	vp->v_flag |= VMOUNT;
	mtx_unlock(&vp->v_interlock);

	/*
	 * Allocate and initialize the filesystem.
	 */
	mp = malloc(sizeof(struct mount), M_MOUNT, M_WAITOK | M_ZERO);
	TAILQ_INIT(&mp->mnt_nvnodelist);
	TAILQ_INIT(&mp->mnt_reservedvnlist);
	lockinit(&mp->mnt_lock, PVFS, "vfslock", 0, LK_NOPAUSE);
	(void)vfs_busy(mp, LK_NOWAIT, 0, td);
	mp->mnt_op = vfsp->vfc_vfsops;
	mp->mnt_vfc = vfsp;
	vfsp->vfc_refcount++;
	mp->mnt_stat.f_type = vfsp->vfc_typenum;
	mp->mnt_flag |= vfsp->vfc_flags & MNT_VISFLAGMASK;
	strncpy(mp->mnt_stat.f_fstypename, fstype, MFSNAMELEN);
	mp->mnt_stat.f_fstypename[MFSNAMELEN - 1] = '\0';
	mp->mnt_vnodecovered = vp;
	mp->mnt_stat.f_owner = td->td_ucred->cr_uid;
	strncpy(mp->mnt_stat.f_mntonname, fspath, MNAMELEN);
	mp->mnt_stat.f_mntonname[MNAMELEN - 1] = '\0';
	mp->mnt_iosize_max = DFLTPHYS;
	VOP_UNLOCK(vp, 0, td);
update:
	/*
	 * Check if the fs implements the old VFS_MOUNT()
	 * function, since the old system call was used.
	 */
	if (mp->mnt_op->vfs_mount == NULL) {
		printf("%s doesn't support the old mount syscall\n",
		    mp->mnt_vfc->vfc_name);
		vput(vp);
		return (EOPNOTSUPP);
	}

	/*
	 * Set the mount level flags.
	 */
	if (fsflags & MNT_RDONLY)
		mp->mnt_flag |= MNT_RDONLY;
	else if (mp->mnt_flag & MNT_RDONLY)
		mp->mnt_kern_flag |= MNTK_WANTRDWR;
	mp->mnt_flag &=~ MNT_UPDATEMASK;
	mp->mnt_flag |= fsflags & (MNT_UPDATEMASK | MNT_FORCE);
	/*
	 * Mount the filesystem.
	 * XXX The final recipients of VFS_MOUNT just overwrite the ndp they
	 * get.  No freeing of cn_pnbuf.
	 */
	error = VFS_MOUNT(mp, fspath, fsdata, &nd, td);
	if (mp->mnt_flag & MNT_UPDATE) {
		if (mp->mnt_kern_flag & MNTK_WANTRDWR)
			mp->mnt_flag &= ~MNT_RDONLY;
		mp->mnt_flag &=~
		    (MNT_UPDATE | MNT_RELOAD | MNT_FORCE | MNT_SNAPSHOT);
		mp->mnt_kern_flag &=~ MNTK_WANTRDWR;
		if (error) {
			mp->mnt_flag = flag;
			mp->mnt_kern_flag = flag2;
		}
		if ((mp->mnt_flag & MNT_RDONLY) == 0) {
			if (mp->mnt_syncer == NULL)
				error = vfs_allocate_syncvnode(mp);
		} else {
			if (mp->mnt_syncer != NULL)
				vrele(mp->mnt_syncer);
			mp->mnt_syncer = NULL;
		}
		vfs_unbusy(mp, td);
		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		mtx_unlock(&vp->v_interlock);
		vrele(vp);
		return (error);
	}
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	/*
	 * Put the new filesystem on the mount list after root.
	 */
	cache_purge(vp);
	if (!error) {
		struct vnode *newdp;

		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		vp->v_mountedhere = mp;
		mtx_unlock(&vp->v_interlock);
		mtx_lock(&mountlist_mtx);
		TAILQ_INSERT_TAIL(&mountlist, mp, mnt_list);
		mtx_unlock(&mountlist_mtx);
		if (VFS_ROOT(mp, &newdp))
			panic("mount: lost mount");
		checkdirs(vp, newdp);
		vput(newdp);
		VOP_UNLOCK(vp, 0, td);
		if ((mp->mnt_flag & MNT_RDONLY) == 0)
			error = vfs_allocate_syncvnode(mp);
		vfs_unbusy(mp, td);
		if ((error = VFS_START(mp, 0, td)) != 0)
			vrele(vp);
	} else {
		mtx_lock(&vp->v_interlock);
		vp->v_flag &= ~VMOUNT;
		mtx_unlock(&vp->v_interlock);
		mp->mnt_vfc->vfc_refcount--;
		vfs_unbusy(mp, td);
		free((caddr_t)mp, M_MOUNT);
		vput(vp);
	}
	return (error);
}

/*
 * Scan all active processes to see if any of them have a current
 * or root directory of `olddp'. If so, replace them with the new
 * mount point.
 */
static void
checkdirs(olddp, newdp)
	struct vnode *olddp, *newdp;
{
	struct filedesc *fdp;
	struct proc *p;
	int nrele;

	if (olddp->v_usecount == 1)
		return;
	sx_slock(&allproc_lock);
	LIST_FOREACH(p, &allproc, p_list) {
		PROC_LOCK(p);
		fdp = p->p_fd;
		if (fdp == NULL) {
			PROC_UNLOCK(p);
			continue;
		}
		nrele = 0;
		FILEDESC_LOCK(fdp);
		if (fdp->fd_cdir == olddp) {
			VREF(newdp);
			fdp->fd_cdir = newdp;
			nrele++;
		}
		if (fdp->fd_rdir == olddp) {
			VREF(newdp);
			fdp->fd_rdir = newdp;
			nrele++;
		}
		FILEDESC_UNLOCK(fdp);
		PROC_UNLOCK(p);
		while (nrele--)
			vrele(olddp);
	}
	sx_sunlock(&allproc_lock);
	if (rootvnode == olddp) {
		vrele(rootvnode);
		VREF(newdp);
		rootvnode = newdp;
	}
}

/*
 * Unmount a file system.
 *
 * Note: unmount takes a path to the vnode mounted on as argument,
 * not special file (as before).
 */
#ifndef _SYS_SYSPROTO_H_
struct unmount_args {
	char	*path;
	int	flags;
};
#endif
/* ARGSUSED */
int
unmount(td, uap)
	struct thread *td;
	register struct unmount_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
	} */ *uap;
{
	register struct vnode *vp;
	struct mount *mp;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	mp = vp->v_mount;

	/*
	 * Only root, or the user that did the original mount is
	 * permitted to unmount this filesystem.
	 */
	if (mp->mnt_stat.f_owner != td->td_ucred->cr_uid) {
		error = suser_td(td);
		if (error) {
			vput(vp);
			return (error);
		}
	}

	/*
	 * Don't allow unmounting the root file system.
	 */
	if (mp->mnt_flag & MNT_ROOTFS) {
		vput(vp);
		return (EINVAL);
	}

	/*
	 * Must be the root of the filesystem
	 */
	if ((vp->v_flag & VROOT) == 0) {
		vput(vp);
		return (EINVAL);
	}
	vput(vp);
	return (dounmount(mp, SCARG(uap, flags), td));
}

/*
 * Do the actual file system unmount.
 */
int
dounmount(mp, flags, td)
	struct mount *mp;
	int flags;
	struct thread *td;
{
	struct vnode *coveredvp, *fsrootvp;
	int error;
	int async_flag;

	mtx_lock(&mountlist_mtx);
	mp->mnt_kern_flag |= MNTK_UNMOUNT;
	error = lockmgr(&mp->mnt_lock, LK_DRAIN | LK_INTERLOCK |
	    ((flags & MNT_FORCE) ? 0 : LK_NOWAIT), &mountlist_mtx, td);
	if (error) {
		mp->mnt_kern_flag &= ~MNTK_UNMOUNT;
		if (mp->mnt_kern_flag & MNTK_MWAIT)
			wakeup((caddr_t)mp);
		return (error);
	}
	vn_start_write(NULL, &mp, V_WAIT);

	if (mp->mnt_flag & MNT_EXPUBLIC)
		vfs_setpublicfs(NULL, NULL, NULL);

	vfs_msync(mp, MNT_WAIT);
	async_flag = mp->mnt_flag & MNT_ASYNC;
	mp->mnt_flag &=~ MNT_ASYNC;
	cache_purgevfs(mp);	/* remove cache entries for this file sys */
	if (mp->mnt_syncer != NULL)
		vrele(mp->mnt_syncer);
	/* Move process cdir/rdir refs on fs root to underlying vnode. */
	if (VFS_ROOT(mp, &fsrootvp) == 0) {
		if (mp->mnt_vnodecovered != NULL)
			checkdirs(fsrootvp, mp->mnt_vnodecovered);
		if (fsrootvp == rootvnode) {
			vrele(rootvnode);
			rootvnode = NULL;
		}
		vput(fsrootvp);
	}
	if (((mp->mnt_flag & MNT_RDONLY) ||
	     (error = VFS_SYNC(mp, MNT_WAIT, td->td_ucred, td)) == 0) ||
	    (flags & MNT_FORCE)) {
		error = VFS_UNMOUNT(mp, flags, td);
	}
	vn_finished_write(mp);
	if (error) {
		/* Undo cdir/rdir and rootvnode changes made above. */
		if (VFS_ROOT(mp, &fsrootvp) == 0) {
			if (mp->mnt_vnodecovered != NULL)
				checkdirs(mp->mnt_vnodecovered, fsrootvp);
			if (rootvnode == NULL) {
				rootvnode = fsrootvp;
				vref(rootvnode);
			}
			vput(fsrootvp);
		}
		if ((mp->mnt_flag & MNT_RDONLY) == 0 && mp->mnt_syncer == NULL)
			(void) vfs_allocate_syncvnode(mp);
		mtx_lock(&mountlist_mtx);
		mp->mnt_kern_flag &= ~MNTK_UNMOUNT;
		mp->mnt_flag |= async_flag;
		lockmgr(&mp->mnt_lock, LK_RELEASE | LK_INTERLOCK,
		    &mountlist_mtx, td);
		if (mp->mnt_kern_flag & MNTK_MWAIT)
			wakeup((caddr_t)mp);
		return (error);
	}
	mtx_lock(&mountlist_mtx);
	TAILQ_REMOVE(&mountlist, mp, mnt_list);
	if ((coveredvp = mp->mnt_vnodecovered) != NULL)
		coveredvp->v_mountedhere = NULL;
	mp->mnt_vfc->vfc_refcount--;
	if (!TAILQ_EMPTY(&mp->mnt_nvnodelist))
		panic("unmount: dangling vnode");
	lockmgr(&mp->mnt_lock, LK_RELEASE | LK_INTERLOCK, &mountlist_mtx, td);
	lockdestroy(&mp->mnt_lock);
	if (coveredvp != NULL)
		vrele(coveredvp);
	if (mp->mnt_kern_flag & MNTK_MWAIT)
		wakeup((caddr_t)mp);
	if (mp->mnt_op->vfs_mount == NULL)
		vfs_freeopts(mp->mnt_opt);
	free((caddr_t)mp, M_MOUNT);
	return (0);
}

/*
 * Sync each mounted filesystem.
 */
#ifndef _SYS_SYSPROTO_H_
struct sync_args {
        int     dummy;
};
#endif

#ifdef DEBUG
static int syncprt = 0;
SYSCTL_INT(_debug, OID_AUTO, syncprt, CTLFLAG_RW, &syncprt, 0, "");
#endif

/* ARGSUSED */
int
sync(td, uap)
	struct thread *td;
	struct sync_args *uap;
{
	struct mount *mp, *nmp;
	int asyncflag;

	mtx_lock(&mountlist_mtx);
	for (mp = TAILQ_FIRST(&mountlist); mp != NULL; mp = nmp) {
		if (vfs_busy(mp, LK_NOWAIT, &mountlist_mtx, td)) {
			nmp = TAILQ_NEXT(mp, mnt_list);
			continue;
		}
		if ((mp->mnt_flag & MNT_RDONLY) == 0 &&
		    vn_start_write(NULL, &mp, V_NOWAIT) == 0) {
			asyncflag = mp->mnt_flag & MNT_ASYNC;
			mp->mnt_flag &= ~MNT_ASYNC;
			vfs_msync(mp, MNT_NOWAIT);
			VFS_SYNC(mp, MNT_NOWAIT,
			    ((td != NULL) ? td->td_ucred : NOCRED), td);
			mp->mnt_flag |= asyncflag;
			vn_finished_write(mp);
		}
		mtx_lock(&mountlist_mtx);
		nmp = TAILQ_NEXT(mp, mnt_list);
		vfs_unbusy(mp, td);
	}
	mtx_unlock(&mountlist_mtx);
#if 0
/*
 * XXX don't call vfs_bufstats() yet because that routine
 * was not imported in the Lite2 merge.
 */
#ifdef DIAGNOSTIC
	if (syncprt)
		vfs_bufstats();
#endif /* DIAGNOSTIC */
#endif
	return (0);
}

/* XXX PRISON: could be per prison flag */
static int prison_quotas;
#if 0
SYSCTL_INT(_kern_prison, OID_AUTO, quotas, CTLFLAG_RW, &prison_quotas, 0, "");
#endif

/*
 * Change filesystem quotas.
 */
#ifndef _SYS_SYSPROTO_H_
struct quotactl_args {
	char *path;
	int cmd;
	int uid;
	caddr_t arg;
};
#endif
/* ARGSUSED */
int
quotactl(td, uap)
	struct thread *td;
	register struct quotactl_args /* {
		syscallarg(char *) path;
		syscallarg(int) cmd;
		syscallarg(int) uid;
		syscallarg(caddr_t) arg;
	} */ *uap;
{
	struct mount *mp;
	int error;
	struct nameidata nd;

	if (jailed(td->td_ucred) && !prison_quotas)
		return (EPERM);
	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = vn_start_write(nd.ni_vp, &mp, V_WAIT | PCATCH);
	vrele(nd.ni_vp);
	if (error)
		return (error);
	error = VFS_QUOTACTL(mp, SCARG(uap, cmd), SCARG(uap, uid),
	    SCARG(uap, arg), td);
	vn_finished_write(mp);
	return (error);
}

/*
 * Get filesystem statistics.
 */
#ifndef _SYS_SYSPROTO_H_
struct statfs_args {
	char *path;
	struct statfs *buf;
};
#endif
/* ARGSUSED */
int
statfs(td, uap)
	struct thread *td;
	register struct statfs_args /* {
		syscallarg(char *) path;
		syscallarg(struct statfs *) buf;
	} */ *uap;
{
	register struct mount *mp;
	register struct statfs *sp;
	int error;
	struct nameidata nd;
	struct statfs sb;

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	mp = nd.ni_vp->v_mount;
	sp = &mp->mnt_stat;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vrele(nd.ni_vp);
	error = VFS_STATFS(mp, sp, td);
	if (error)
		return (error);
	sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
	if (suser_xxx(td->td_ucred, 0, 0)) {
		bcopy((caddr_t)sp, (caddr_t)&sb, sizeof(sb));
		sb.f_fsid.val[0] = sb.f_fsid.val[1] = 0;
		sp = &sb;
	}
	return (copyout((caddr_t)sp, (caddr_t)SCARG(uap, buf), sizeof(*sp)));
}

/*
 * Get filesystem statistics.
 */
#ifndef _SYS_SYSPROTO_H_
struct fstatfs_args {
	int fd;
	struct statfs *buf;
};
#endif
/* ARGSUSED */
int
fstatfs(td, uap)
	struct thread *td;
	register struct fstatfs_args /* {
		syscallarg(int) fd;
		syscallarg(struct statfs *) buf;
	} */ *uap;
{
	struct file *fp;
	struct mount *mp;
	register struct statfs *sp;
	int error;
	struct statfs sb;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	mp = ((struct vnode *)fp->f_data)->v_mount;
	fdrop(fp, td);
	if (mp == NULL)
		return (EBADF);
	sp = &mp->mnt_stat;
	error = VFS_STATFS(mp, sp, td);
	if (error)
		return (error);
	sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
	if (suser_xxx(td->td_ucred, 0, 0)) {
		bcopy((caddr_t)sp, (caddr_t)&sb, sizeof(sb));
		sb.f_fsid.val[0] = sb.f_fsid.val[1] = 0;
		sp = &sb;
	}
	return (copyout((caddr_t)sp, (caddr_t)SCARG(uap, buf), sizeof(*sp)));
}

/*
 * Get statistics on all filesystems.
 */
#ifndef _SYS_SYSPROTO_H_
struct getfsstat_args {
	struct statfs *buf;
	long bufsize;
	int flags;
};
#endif
int
getfsstat(td, uap)
	struct thread *td;
	register struct getfsstat_args /* {
		syscallarg(struct statfs *) buf;
		syscallarg(long) bufsize;
		syscallarg(int) flags;
	} */ *uap;
{
	register struct mount *mp, *nmp;
	register struct statfs *sp;
	caddr_t sfsp;
	long count, maxcount, error;

	maxcount = SCARG(uap, bufsize) / sizeof(struct statfs);
	sfsp = (caddr_t)SCARG(uap, buf);
	count = 0;
	mtx_lock(&mountlist_mtx);
	for (mp = TAILQ_FIRST(&mountlist); mp != NULL; mp = nmp) {
		if (vfs_busy(mp, LK_NOWAIT, &mountlist_mtx, td)) {
			nmp = TAILQ_NEXT(mp, mnt_list);
			continue;
		}
		if (sfsp && count < maxcount) {
			sp = &mp->mnt_stat;
			/*
			 * If MNT_NOWAIT or MNT_LAZY is specified, do not
			 * refresh the fsstat cache. MNT_NOWAIT or MNT_LAZY
			 * overrides MNT_WAIT.
			 */
			if (((SCARG(uap, flags) & (MNT_LAZY|MNT_NOWAIT)) == 0 ||
			    (SCARG(uap, flags) & MNT_WAIT)) &&
			    (error = VFS_STATFS(mp, sp, td))) {
				mtx_lock(&mountlist_mtx);
				nmp = TAILQ_NEXT(mp, mnt_list);
				vfs_unbusy(mp, td);
				continue;
			}
			sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
			error = copyout((caddr_t)sp, sfsp, sizeof(*sp));
			if (error) {
				vfs_unbusy(mp, td);
				return (error);
			}
			sfsp += sizeof(*sp);
		}
		count++;
		mtx_lock(&mountlist_mtx);
		nmp = TAILQ_NEXT(mp, mnt_list);
		vfs_unbusy(mp, td);
	}
	mtx_unlock(&mountlist_mtx);
	if (sfsp && count > maxcount)
		td->td_retval[0] = maxcount;
	else
		td->td_retval[0] = count;
	return (0);
}

/*
 * Change current working directory to a given file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct fchdir_args {
	int	fd;
};
#endif
/* ARGSUSED */
int
fchdir(td, uap)
	struct thread *td;
	struct fchdir_args /* {
		syscallarg(int) fd;
	} */ *uap;
{
	register struct filedesc *fdp = td->td_proc->p_fd;
	struct vnode *vp, *tdp, *vpold;
	struct mount *mp;
	struct file *fp;
	int error;

	if ((error = getvnode(fdp, SCARG(uap, fd), &fp)) != 0)
		return (error);
	vp = (struct vnode *)fp->f_data;
	VREF(vp);
	fdrop(fp, td);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	if (vp->v_type != VDIR)
		error = ENOTDIR;
	else
		error = VOP_ACCESS(vp, VEXEC, td->td_ucred, td);
	while (!error && (mp = vp->v_mountedhere) != NULL) {
		if (vfs_busy(mp, 0, 0, td))
			continue;
		error = VFS_ROOT(mp, &tdp);
		vfs_unbusy(mp, td);
		if (error)
			break;
		vput(vp);
		vp = tdp;
	}
	if (error) {
		vput(vp);
		return (error);
	}
	VOP_UNLOCK(vp, 0, td);
	FILEDESC_LOCK(fdp);
	vpold = fdp->fd_cdir;
	fdp->fd_cdir = vp;
	FILEDESC_UNLOCK(fdp);
	vrele(vpold);
	return (0);
}

/*
 * Change current working directory (``.'').
 */
#ifndef _SYS_SYSPROTO_H_
struct chdir_args {
	char	*path;
};
#endif
/* ARGSUSED */
int
chdir(td, uap)
	struct thread *td;
	struct chdir_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	register struct filedesc *fdp = td->td_proc->p_fd;
	int error;
	struct nameidata nd;
	struct vnode *vp;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = change_dir(&nd, td)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	FILEDESC_LOCK(fdp);
	vp = fdp->fd_cdir;
	fdp->fd_cdir = nd.ni_vp;
	FILEDESC_UNLOCK(fdp);
	vrele(vp);
	return (0);
}

/*
 * Helper function for raised chroot(2) security function:  Refuse if
 * any filedescriptors are open directories.
 */
static int
chroot_refuse_vdir_fds(fdp)
	struct filedesc *fdp;
{
	struct vnode *vp;
	struct file *fp;
	int fd;

	FILEDESC_LOCK(fdp);
	for (fd = 0; fd < fdp->fd_nfiles ; fd++) {
		fp = fget_locked(fdp, fd);
		if (fp == NULL)
			continue;
		if (fp->f_type == DTYPE_VNODE) {
			vp = (struct vnode *)fp->f_data;
			if (vp->v_type == VDIR) {
				FILEDESC_UNLOCK(fdp);
				return (EPERM);
			}
		}
	}
	FILEDESC_UNLOCK(fdp);
	return (0);
}

/*
 * This sysctl determines if we will allow a process to chroot(2) if it
 * has a directory open:
 *	0: disallowed for all processes.
 *	1: allowed for processes that were not already chroot(2)'ed.
 *	2: allowed for all processes.
 */

static int chroot_allow_open_directories = 1;

SYSCTL_INT(_kern, OID_AUTO, chroot_allow_open_directories, CTLFLAG_RW,
     &chroot_allow_open_directories, 0, "");

/*
 * Change notion of root (``/'') directory.
 */
#ifndef _SYS_SYSPROTO_H_
struct chroot_args {
	char	*path;
};
#endif
/* ARGSUSED */
int
chroot(td, uap)
	struct thread *td;
	struct chroot_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	register struct filedesc *fdp = td->td_proc->p_fd;
	int error;
	struct nameidata nd;
	struct vnode *vp;

	error = suser_xxx(0, td->td_proc, PRISON_ROOT);
	if (error)
		return (error);
	FILEDESC_LOCK(fdp);
	if (chroot_allow_open_directories == 0 ||
	    (chroot_allow_open_directories == 1 && fdp->fd_rdir != rootvnode)) {
		FILEDESC_UNLOCK(fdp);
		error = chroot_refuse_vdir_fds(fdp);
	} else
		FILEDESC_UNLOCK(fdp);
	if (error)
		return (error);
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = change_dir(&nd, td)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	FILEDESC_LOCK(fdp);
	vp = fdp->fd_rdir;
	fdp->fd_rdir = nd.ni_vp;
	if (!fdp->fd_jdir) {
		fdp->fd_jdir = nd.ni_vp;
                VREF(fdp->fd_jdir);
	}
	FILEDESC_UNLOCK(fdp);
	vrele(vp);
	return (0);
}

/*
 * Common routine for chroot and chdir.
 */
static int
change_dir(ndp, td)
	register struct nameidata *ndp;
	struct thread *td;
{
	struct vnode *vp;
	int error;

	error = namei(ndp);
	if (error)
		return (error);
	vp = ndp->ni_vp;
	if (vp->v_type != VDIR)
		error = ENOTDIR;
	else
		error = VOP_ACCESS(vp, VEXEC, td->td_ucred, td);
	if (error)
		vput(vp);
	else
		VOP_UNLOCK(vp, 0, td);
	return (error);
}

/*
 * Check permissions, allocate an open file structure,
 * and call the device open routine if any.
 */
#ifndef _SYS_SYSPROTO_H_
struct open_args {
	char	*path;
	int	flags;
	int	mode;
};
#endif
int
open(td, uap)
	struct thread *td;
	register struct open_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
		syscallarg(int) mode;
	} */ *uap;
{
	struct proc *p = td->td_proc;
	struct filedesc *fdp = p->p_fd;
	struct file *fp;
	struct vnode *vp;
	struct vattr vat;
	struct mount *mp;
	int cmode, flags, oflags;
	struct file *nfp;
	int type, indx, error;
	struct flock lf;
	struct nameidata nd;

	oflags = SCARG(uap, flags);
	if ((oflags & O_ACCMODE) == O_ACCMODE)
		return (EINVAL);
	flags = FFLAGS(oflags);
	error = falloc(td, &nfp, &indx);
	if (error)
		return (error);
	fp = nfp;
	FILEDESC_LOCK(fdp);
	cmode = ((SCARG(uap, mode) &~ fdp->fd_cmask) & ALLPERMS) &~ S_ISTXT;
	FILEDESC_UNLOCK(fdp);
	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	td->td_dupfd = -indx - 1;		/* XXX check for fdopen */
	/*
	 * Bump the ref count to prevent another process from closing
	 * the descriptor while we are blocked in vn_open()
	 */
	fhold(fp);
	error = vn_open(&nd, &flags, cmode);
	if (error) {
		/*
		 * release our own reference
		 */
		fdrop(fp, td);

		/*
		 * handle special fdopen() case.  bleh.  dupfdopen() is
		 * responsible for dropping the old contents of ofiles[indx]
		 * if it succeeds.
		 */
		if ((error == ENODEV || error == ENXIO) &&
		    td->td_dupfd >= 0 &&		/* XXX from fdopen */
		    (error =
			dupfdopen(td, fdp, indx, td->td_dupfd, flags, error)) == 0) {
			td->td_retval[0] = indx;
			return (0);
		}
		/*
		 * Clean up the descriptor, but only if another thread hadn't
		 * replaced or closed it.
		 */
		FILEDESC_LOCK(fdp);
		if (fdp->fd_ofiles[indx] == fp) {
			fdp->fd_ofiles[indx] = NULL;
			FILEDESC_UNLOCK(fdp);
			fdrop(fp, td);
		} else
			FILEDESC_UNLOCK(fdp);

		if (error == ERESTART)
			error = EINTR;
		return (error);
	}
	td->td_dupfd = 0;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;

	/*
	 * There should be 2 references on the file, one from the descriptor
	 * table, and one for us.
	 *
	 * Handle the case where someone closed the file (via its file
	 * descriptor) while we were blocked.  The end result should look
	 * like opening the file succeeded but it was immediately closed.
	 */
	FILEDESC_LOCK(fdp);
	FILE_LOCK(fp);
	if (fp->f_count == 1) {
		KASSERT(fdp->fd_ofiles[indx] != fp,
		    ("Open file descriptor lost all refs"));
		FILEDESC_UNLOCK(fdp);
		FILE_UNLOCK(fp);
		VOP_UNLOCK(vp, 0, td);
		vn_close(vp, flags & FMASK, fp->f_cred, td);
		fdrop(fp, td);
		td->td_retval[0] = indx;
		return 0;
	}

	fp->f_data = (caddr_t)vp;
	fp->f_flag = flags & FMASK;
	fp->f_ops = &vnops;
	fp->f_type = (vp->v_type == VFIFO ? DTYPE_FIFO : DTYPE_VNODE);
	FILEDESC_UNLOCK(fdp);
	FILE_UNLOCK(fp);
	VOP_UNLOCK(vp, 0, td);
	if (flags & (O_EXLOCK | O_SHLOCK)) {
		lf.l_whence = SEEK_SET;
		lf.l_start = 0;
		lf.l_len = 0;
		if (flags & O_EXLOCK)
			lf.l_type = F_WRLCK;
		else
			lf.l_type = F_RDLCK;
		type = F_FLOCK;
		if ((flags & FNONBLOCK) == 0)
			type |= F_WAIT;
		if ((error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type)) != 0)
			goto bad;
		fp->f_flag |= FHASLOCK;
	}
	if (flags & O_TRUNC) {
		if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
			goto bad;
		VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
		VATTR_NULL(&vat);
		vat.va_size = 0;
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
		error = VOP_SETATTR(vp, &vat, td->td_ucred, td);
		VOP_UNLOCK(vp, 0, td);
		vn_finished_write(mp);
		if (error)
			goto bad;
	}
	/* assert that vn_open created a backing object if one is needed */
	KASSERT(!vn_canvmio(vp) || VOP_GETVOBJECT(vp, NULL) == 0,
		("open: vmio vnode has no backing object after vn_open"));
	/*
	 * Release our private reference, leaving the one associated with
	 * the descriptor table intact.
	 */
	fdrop(fp, td);
	td->td_retval[0] = indx;
	return (0);
bad:
	FILEDESC_LOCK(fdp);
	if (fdp->fd_ofiles[indx] == fp) {
		fdp->fd_ofiles[indx] = NULL;
		FILEDESC_UNLOCK(fdp);
		fdrop(fp, td);
	} else
		FILEDESC_UNLOCK(fdp);
	return (error);
}

#ifdef COMPAT_43
/*
 * Create a file.
 */
#ifndef _SYS_SYSPROTO_H_
struct ocreat_args {
	char	*path;
	int	mode;
};
#endif
int
ocreat(td, uap)
	struct thread *td;
	register struct ocreat_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
	} */ *uap;
{
	struct open_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
		syscallarg(int) mode;
	} */ nuap;

	SCARG(&nuap, path) = SCARG(uap, path);
	SCARG(&nuap, mode) = SCARG(uap, mode);
	SCARG(&nuap, flags) = O_WRONLY | O_CREAT | O_TRUNC;
	return (open(td, &nuap));
}
#endif /* COMPAT_43 */

/*
 * Create a special file.
 */
#ifndef _SYS_SYSPROTO_H_
struct mknod_args {
	char	*path;
	int	mode;
	int	dev;
};
#endif
/* ARGSUSED */
int
mknod(td, uap)
	struct thread *td;
	register struct mknod_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
		syscallarg(int) dev;
	} */ *uap;
{
	struct vnode *vp;
	struct mount *mp;
	struct vattr vattr;
	int error;
	int whiteout = 0;
	struct nameidata nd;

	switch (SCARG(uap, mode) & S_IFMT) {
	case S_IFCHR:
	case S_IFBLK:
		error = suser_td(td);
		break;
	default:
		error = suser_xxx(0, td->td_proc, PRISON_ROOT);
		break;
	}
	if (error)
		return (error);
restart:
	bwillwrite();
	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	if (vp != NULL) {
		vrele(vp);
		error = EEXIST;
	} else {
		VATTR_NULL(&vattr);
		FILEDESC_LOCK(td->td_proc->p_fd);
		vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ td->td_proc->p_fd->fd_cmask;
		FILEDESC_UNLOCK(td->td_proc->p_fd);
		vattr.va_rdev = SCARG(uap, dev);
		whiteout = 0;

		switch (SCARG(uap, mode) & S_IFMT) {
		case S_IFMT:	/* used by badsect to flag bad sectors */
			vattr.va_type = VBAD;
			break;
		case S_IFCHR:
			vattr.va_type = VCHR;
			break;
		case S_IFBLK:
			vattr.va_type = VBLK;
			break;
		case S_IFWHT:
			whiteout = 1;
			break;
		default:
			error = EINVAL;
			break;
		}
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	if (!error) {
		VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
		if (whiteout)
			error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, CREATE);
		else {
			error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp,
						&nd.ni_cnd, &vattr);
			if (error == 0)
				vput(nd.ni_vp);
		}
	}
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_dvp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "mknod");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "mknod");
	return (error);
}

/*
 * Create a named pipe.
 */
#ifndef _SYS_SYSPROTO_H_
struct mkfifo_args {
	char	*path;
	int	mode;
};
#endif
/* ARGSUSED */
int
mkfifo(td, uap)
	struct thread *td;
	register struct mkfifo_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
	} */ *uap;
{
	struct mount *mp;
	struct vattr vattr;
	int error;
	struct nameidata nd;

restart:
	bwillwrite();
	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	if (nd.ni_vp != NULL) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vrele(nd.ni_vp);
		vput(nd.ni_dvp);
		return (EEXIST);
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VATTR_NULL(&vattr);
	vattr.va_type = VFIFO;
	FILEDESC_LOCK(td->td_proc->p_fd);
	vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ td->td_proc->p_fd->fd_cmask;
	FILEDESC_UNLOCK(td->td_proc->p_fd);
	VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
	error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr);
	if (error == 0)
		vput(nd.ni_vp);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_dvp);
	vn_finished_write(mp);
	return (error);
}

/*
 * Make a hard file link.
 */
#ifndef _SYS_SYSPROTO_H_
struct link_args {
	char	*path;
	char	*link;
};
#endif
/* ARGSUSED */
int
link(td, uap)
	struct thread *td;
	register struct link_args /* {
		syscallarg(char *) path;
		syscallarg(char *) link;
	} */ *uap;
{
	struct vnode *vp;
	struct mount *mp;
	struct nameidata nd;
	int error;

	bwillwrite();
	NDINIT(&nd, LOOKUP, FOLLOW|NOOBJ, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;
	if (vp->v_type == VDIR) {
		vrele(vp);
		return (EPERM);		/* POSIX */
	}
	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) {
		vrele(vp);
		return (error);
	}
	NDINIT(&nd, CREATE, LOCKPARENT|NOOBJ, UIO_USERSPACE, SCARG(uap, link), td);
	if ((error = namei(&nd)) == 0) {
		if (nd.ni_vp != NULL) {
			vrele(nd.ni_vp);
			error = EEXIST;
		} else {
			VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
			VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
			error = VOP_LINK(nd.ni_dvp, vp, &nd.ni_cnd);
		}
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
	}
	vrele(vp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "link");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "link");
	return (error);
}

/*
 * Make a symbolic link.
 */
#ifndef _SYS_SYSPROTO_H_
struct symlink_args {
	char	*path;
	char	*link;
};
#endif
/* ARGSUSED */
int
symlink(td, uap)
	struct thread *td;
	register struct symlink_args /* {
		syscallarg(char *) path;
		syscallarg(char *) link;
	} */ *uap;
{
	struct mount *mp;
	struct vattr vattr;
	char *path;
	int error;
	struct nameidata nd;

	path = uma_zalloc(namei_zone, M_WAITOK);
	if ((error = copyinstr(SCARG(uap, path), path, MAXPATHLEN, NULL)) != 0)
		goto out;
restart:
	bwillwrite();
	NDINIT(&nd, CREATE, LOCKPARENT|NOOBJ, UIO_USERSPACE, SCARG(uap, link), td);
	if ((error = namei(&nd)) != 0)
		goto out;
	if (nd.ni_vp) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vrele(nd.ni_vp);
		vput(nd.ni_dvp);
		error = EEXIST;
		goto out;
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VATTR_NULL(&vattr);
	FILEDESC_LOCK(td->td_proc->p_fd);
	vattr.va_mode = ACCESSPERMS &~ td->td_proc->p_fd->fd_cmask;
	FILEDESC_UNLOCK(td->td_proc->p_fd);
	VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
	error = VOP_SYMLINK(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr, path);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	if (error == 0)
		vput(nd.ni_vp);
	vput(nd.ni_dvp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "symlink");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "symlink");
out:
	uma_zfree(namei_zone, path);
	return (error);
}

/*
 * Delete a whiteout from the filesystem.
 */
/* ARGSUSED */
int
undelete(td, uap)
	struct thread *td;
	register struct undelete_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	int error;
	struct mount *mp;
	struct nameidata nd;

restart:
	bwillwrite();
	NDINIT(&nd, DELETE, LOCKPARENT|DOWHITEOUT, UIO_USERSPACE,
	    SCARG(uap, path), td);
	error = namei(&nd);
	if (error)
		return (error);

	if (nd.ni_vp != NULLVP || !(nd.ni_cnd.cn_flags & ISWHITEOUT)) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		if (nd.ni_vp)
			vrele(nd.ni_vp);
		vput(nd.ni_dvp);
		return (EEXIST);
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
	error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, DELETE);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_dvp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "undelete");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "undelete");
	return (error);
}

/*
 * Delete a name from the filesystem.
 */
#ifndef _SYS_SYSPROTO_H_
struct unlink_args {
	char	*path;
};
#endif
/* ARGSUSED */
int
unlink(td, uap)
	struct thread *td;
	struct unlink_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	struct mount *mp;
	struct vnode *vp;
	int error;
	struct nameidata nd;

restart:
	bwillwrite();
	NDINIT(&nd, DELETE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	if (vp->v_type == VDIR)
		error = EPERM;		/* POSIX */
	else {
		/*
		 * The root of a mounted filesystem cannot be deleted.
		 *
		 * XXX: can this only be a VDIR case?
		 */
		if (vp->v_flag & VROOT)
			error = EBUSY;
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vrele(vp);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	if (!error) {
		VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
		error = VOP_REMOVE(nd.ni_dvp, vp, &nd.ni_cnd);
	}
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_dvp);
	vput(vp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "unlink");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "unlink");
	return (error);
}

/*
 * Reposition read/write file offset.
 */
#ifndef _SYS_SYSPROTO_H_
struct lseek_args {
	int	fd;
	int	pad;
	off_t	offset;
	int	whence;
};
#endif
int
lseek(td, uap)
	struct thread *td;
	register struct lseek_args /* {
		syscallarg(int) fd;
		syscallarg(int) pad;
		syscallarg(off_t) offset;
		syscallarg(int) whence;
	} */ *uap;
{
	struct ucred *cred = td->td_ucred;
	struct file *fp;
	struct vnode *vp;
	struct vattr vattr;
	off_t offset;
	int error, noneg;

	if ((error = fget(td, uap->fd, &fp)) != 0)
		return (error);
	if (fp->f_type != DTYPE_VNODE) {
		fdrop(fp, td);
		return (ESPIPE);
	}
	vp = (struct vnode *)fp->f_data;
	noneg = (vp->v_type != VCHR);
	offset = SCARG(uap, offset);
	switch (SCARG(uap, whence)) {
	case L_INCR:
		if (noneg &&
		    (fp->f_offset < 0 ||
		     (offset > 0 && fp->f_offset > OFF_MAX - offset)))
			return (EOVERFLOW);
		offset += fp->f_offset;
		break;
	case L_XTND:
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
		error = VOP_GETATTR(vp, &vattr, cred, td);
		VOP_UNLOCK(vp, 0, td);
		if (error)
			return (error);
		if (noneg &&
		    (vattr.va_size > OFF_MAX ||
		     (offset > 0 && vattr.va_size > OFF_MAX - offset)))
			return (EOVERFLOW);
		offset += vattr.va_size;
		break;
	case L_SET:
		break;
	default:
		fdrop(fp, td);
		return (EINVAL);
	}
	if (noneg && offset < 0)
		return (EINVAL);
	fp->f_offset = offset;
	*(off_t *)(td->td_retval) = fp->f_offset;
	fdrop(fp, td);
	return (0);
}

#if defined(COMPAT_43) || defined(COMPAT_SUNOS)
/*
 * Reposition read/write file offset.
 */
#ifndef _SYS_SYSPROTO_H_
struct olseek_args {
	int	fd;
	long	offset;
	int	whence;
};
#endif
int
olseek(td, uap)
	struct thread *td;
	register struct olseek_args /* {
		syscallarg(int) fd;
		syscallarg(long) offset;
		syscallarg(int) whence;
	} */ *uap;
{
	struct lseek_args /* {
		syscallarg(int) fd;
		syscallarg(int) pad;
		syscallarg(off_t) offset;
		syscallarg(int) whence;
	} */ nuap;
	int error;

	SCARG(&nuap, fd) = SCARG(uap, fd);
	SCARG(&nuap, offset) = SCARG(uap, offset);
	SCARG(&nuap, whence) = SCARG(uap, whence);
	error = lseek(td, &nuap);
	return (error);
}
#endif /* COMPAT_43 */

/*
 * Check access permissions using passed credentials.
 */
static int
vn_access(vp, user_flags, cred, td)
	struct vnode	*vp;
	int		user_flags;
	struct ucred	*cred;
	struct thread	*td;
{
	int error, flags;

	/* Flags == 0 means only check for existence. */
	error = 0;
	if (user_flags) {
		flags = 0;
		if (user_flags & R_OK)
			flags |= VREAD;
		if (user_flags & W_OK)
			flags |= VWRITE;
		if (user_flags & X_OK)
			flags |= VEXEC;
		if ((flags & VWRITE) == 0 || (error = vn_writechk(vp)) == 0)
			error = VOP_ACCESS(vp, flags, cred, td);
	}
	return (error);
}

/*
 * Check access permissions using "real" credentials.
 */
#ifndef _SYS_SYSPROTO_H_
struct access_args {
	char	*path;
	int	flags;
};
#endif
int
access(td, uap)
	struct thread *td;
	register struct access_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
	} */ *uap;
{
	struct ucred *cred, *tmpcred;
	register struct vnode *vp;
	int error;
	struct nameidata nd;

	/*
	 * Create and modify a temporary credential instead of one that
	 * is potentially shared.  This could also mess up socket
	 * buffer accounting which can run in an interrupt context.
	 *
	 * XXX - Depending on how "threads" are finally implemented, it
	 * may be better to explicitly pass the credential to namei()
	 * rather than to modify the potentially shared process structure.
	 */
	cred = td->td_ucred;
	tmpcred = crdup(cred);
	tmpcred->cr_uid = cred->cr_ruid;
	tmpcred->cr_groups[0] = cred->cr_rgid;
	td->td_ucred = tmpcred;
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		goto out1;
	vp = nd.ni_vp;

	error = vn_access(vp, SCARG(uap, flags), tmpcred, td);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(vp);
out1:
	td->td_ucred = cred;
	crfree(tmpcred);
	return (error);
}

/*
 * Check access permissions using "effective" credentials.
 */
#ifndef _SYS_SYSPROTO_H_
struct eaccess_args {
	char	*path;
	int	flags;
};
#endif
int
eaccess(td, uap)
	struct thread *td;
	register struct eaccess_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
	} */ *uap;
{
	struct nameidata nd;
	struct vnode *vp;
	int error;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;

	error = vn_access(vp, SCARG(uap, flags), td->td_ucred, td);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(vp);
	return (error);
}

#if defined(COMPAT_43) || defined(COMPAT_SUNOS)
/*
 * Get file status; this version follows links.
 */
#ifndef _SYS_SYSPROTO_H_
struct ostat_args {
	char	*path;
	struct ostat *ub;
};
#endif
/* ARGSUSED */
int
ostat(td, uap)
	struct thread *td;
	register struct ostat_args /* {
		syscallarg(char *) path;
		syscallarg(struct ostat *) ub;
	} */ *uap;
{
	struct stat sb;
	struct ostat osb;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = vn_stat(nd.ni_vp, &sb, td);
	vput(nd.ni_vp);
	if (error)
		return (error);
	cvtstat(&sb, &osb);
	error = copyout((caddr_t)&osb, (caddr_t)SCARG(uap, ub), sizeof (osb));
	return (error);
}

/*
 * Get file status; this version does not follow links.
 */
#ifndef _SYS_SYSPROTO_H_
struct olstat_args {
	char	*path;
	struct ostat *ub;
};
#endif
/* ARGSUSED */
int
olstat(td, uap)
	struct thread *td;
	register struct olstat_args /* {
		syscallarg(char *) path;
		syscallarg(struct ostat *) ub;
	} */ *uap;
{
	struct vnode *vp;
	struct stat sb;
	struct ostat osb;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	error = vn_stat(vp, &sb, td);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(vp);
	if (error)
		return (error);
	cvtstat(&sb, &osb);
	error = copyout((caddr_t)&osb, (caddr_t)SCARG(uap, ub), sizeof (osb));
	return (error);
}

/*
 * Convert from an old to a new stat structure.
 */
void
cvtstat(st, ost)
	struct stat *st;
	struct ostat *ost;
{

	ost->st_dev = st->st_dev;
	ost->st_ino = st->st_ino;
	ost->st_mode = st->st_mode;
	ost->st_nlink = st->st_nlink;
	ost->st_uid = st->st_uid;
	ost->st_gid = st->st_gid;
	ost->st_rdev = st->st_rdev;
	if (st->st_size < (quad_t)1 << 32)
		ost->st_size = st->st_size;
	else
		ost->st_size = -2;
	ost->st_atime = st->st_atime;
	ost->st_mtime = st->st_mtime;
	ost->st_ctime = st->st_ctime;
	ost->st_blksize = st->st_blksize;
	ost->st_blocks = st->st_blocks;
	ost->st_flags = st->st_flags;
	ost->st_gen = st->st_gen;
}
#endif /* COMPAT_43 || COMPAT_SUNOS */

/*
 * Get file status; this version follows links.
 */
#ifndef _SYS_SYSPROTO_H_
struct stat_args {
	char	*path;
	struct stat *ub;
};
#endif
/* ARGSUSED */
int
stat(td, uap)
	struct thread *td;
	register struct stat_args /* {
		syscallarg(char *) path;
		syscallarg(struct stat *) ub;
	} */ *uap;
{
	struct stat sb;
	int error;
	struct nameidata nd;

#ifdef LOOKUP_SHARED
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKSHARED | LOCKLEAF | NOOBJ,
	    UIO_USERSPACE, SCARG(uap, path), td);
#else
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
#endif
	if ((error = namei(&nd)) != 0)
		return (error);
	error = vn_stat(nd.ni_vp, &sb, td);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_vp);
	if (error)
		return (error);
	error = copyout((caddr_t)&sb, (caddr_t)SCARG(uap, ub), sizeof (sb));
	return (error);
}

/*
 * Get file status; this version does not follow links.
 */
#ifndef _SYS_SYSPROTO_H_
struct lstat_args {
	char	*path;
	struct stat *ub;
};
#endif
/* ARGSUSED */
int
lstat(td, uap)
	struct thread *td;
	register struct lstat_args /* {
		syscallarg(char *) path;
		syscallarg(struct stat *) ub;
	} */ *uap;
{
	int error;
	struct vnode *vp;
	struct stat sb;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	error = vn_stat(vp, &sb, td);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(vp);
	if (error)
		return (error);
	error = copyout((caddr_t)&sb, (caddr_t)SCARG(uap, ub), sizeof (sb));
	return (error);
}

/*
 * Implementation of the NetBSD stat() function.
 * XXX This should probably be collapsed with the FreeBSD version,
 * as the differences are only due to vn_stat() clearing spares at
 * the end of the structures.  vn_stat could be split to avoid this,
 * and thus collapse the following to close to zero code.
 */
void
cvtnstat(sb, nsb)
	struct stat *sb;
	struct nstat *nsb;
{
	nsb->st_dev = sb->st_dev;
	nsb->st_ino = sb->st_ino;
	nsb->st_mode = sb->st_mode;
	nsb->st_nlink = sb->st_nlink;
	nsb->st_uid = sb->st_uid;
	nsb->st_gid = sb->st_gid;
	nsb->st_rdev = sb->st_rdev;
	nsb->st_atimespec = sb->st_atimespec;
	nsb->st_mtimespec = sb->st_mtimespec;
	nsb->st_ctimespec = sb->st_ctimespec;
	nsb->st_size = sb->st_size;
	nsb->st_blocks = sb->st_blocks;
	nsb->st_blksize = sb->st_blksize;
	nsb->st_flags = sb->st_flags;
	nsb->st_gen = sb->st_gen;
	nsb->st_qspare[0] = sb->st_qspare[0];
	nsb->st_qspare[1] = sb->st_qspare[1];
}

#ifndef _SYS_SYSPROTO_H_
struct nstat_args {
	char	*path;
	struct nstat *ub;
};
#endif
/* ARGSUSED */
int
nstat(td, uap)
	struct thread *td;
	register struct nstat_args /* {
		syscallarg(char *) path;
		syscallarg(struct nstat *) ub;
	} */ *uap;
{
	struct stat sb;
	struct nstat nsb;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = vn_stat(nd.ni_vp, &sb, td);
	vput(nd.ni_vp);
	if (error)
		return (error);
	cvtnstat(&sb, &nsb);
	error = copyout((caddr_t)&nsb, (caddr_t)SCARG(uap, ub), sizeof (nsb));
	return (error);
}

/*
 * NetBSD lstat.  Get file status; this version does not follow links.
 */
#ifndef _SYS_SYSPROTO_H_
struct lstat_args {
	char	*path;
	struct stat *ub;
};
#endif
/* ARGSUSED */
int
nlstat(td, uap)
	struct thread *td;
	register struct nlstat_args /* {
		syscallarg(char *) path;
		syscallarg(struct nstat *) ub;
	} */ *uap;
{
	int error;
	struct vnode *vp;
	struct stat sb;
	struct nstat nsb;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = vn_stat(vp, &sb, td);
	vput(vp);
	if (error)
		return (error);
	cvtnstat(&sb, &nsb);
	error = copyout((caddr_t)&nsb, (caddr_t)SCARG(uap, ub), sizeof (nsb));
	return (error);
}

/*
 * Get configurable pathname variables.
 */
#ifndef _SYS_SYSPROTO_H_
struct pathconf_args {
	char	*path;
	int	name;
};
#endif
/* ARGSUSED */
int
pathconf(td, uap)
	struct thread *td;
	register struct pathconf_args /* {
		syscallarg(char *) path;
		syscallarg(int) name;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = VOP_PATHCONF(nd.ni_vp, SCARG(uap, name), td->td_retval);
	vput(nd.ni_vp);
	return (error);
}

/*
 * Return target name of a symbolic link.
 */
#ifndef _SYS_SYSPROTO_H_
struct readlink_args {
	char	*path;
	char	*buf;
	int	count;
};
#endif
/* ARGSUSED */
int
readlink(td, uap)
	struct thread *td;
	register struct readlink_args /* {
		syscallarg(char *) path;
		syscallarg(char *) buf;
		syscallarg(int) count;
	} */ *uap;
{
	register struct vnode *vp;
	struct iovec aiov;
	struct uio auio;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | NOOBJ, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;
	if (vp->v_type != VLNK)
		error = EINVAL;
	else {
		aiov.iov_base = SCARG(uap, buf);
		aiov.iov_len = SCARG(uap, count);
		auio.uio_iov = &aiov;
		auio.uio_iovcnt = 1;
		auio.uio_offset = 0;
		auio.uio_rw = UIO_READ;
		auio.uio_segflg = UIO_USERSPACE;
		auio.uio_td = td;
		auio.uio_resid = SCARG(uap, count);
		error = VOP_READLINK(vp, &auio, td->td_ucred);
	}
	vput(vp);
	td->td_retval[0] = SCARG(uap, count) - auio.uio_resid;
	return (error);
}

/*
 * Common implementation code for chflags() and fchflags().
 */
static int
setfflags(td, vp, flags)
	struct thread *td;
	struct vnode *vp;
	int flags;
{
	int error;
	struct mount *mp;
	struct vattr vattr;

	/*
	 * Prevent non-root users from setting flags on devices.  When
	 * a device is reused, users can retain ownership of the device
	 * if they are allowed to set flags and programs assume that
	 * chown can't fail when done as root.
	 */
	if (vp->v_type == VCHR || vp->v_type == VBLK) {
		error = suser_xxx(td->td_ucred, td->td_proc, PRISON_ROOT);
		if (error)
			return (error);
	}

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	VATTR_NULL(&vattr);
	vattr.va_flags = flags;
	error = VOP_SETATTR(vp, &vattr, td->td_ucred, td);
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return (error);
}

/*
 * Change flags of a file given a path name.
 */
#ifndef _SYS_SYSPROTO_H_
struct chflags_args {
	char	*path;
	int	flags;
};
#endif
/* ARGSUSED */
int
chflags(td, uap)
	struct thread *td;
	register struct chflags_args /* {
		syscallarg(char *) path;
		syscallarg(int) flags;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setfflags(td, nd.ni_vp, SCARG(uap, flags));
	vrele(nd.ni_vp);
	return error;
}

/*
 * Change flags of a file given a file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct fchflags_args {
	int	fd;
	int	flags;
};
#endif
/* ARGSUSED */
int
fchflags(td, uap)
	struct thread *td;
	register struct fchflags_args /* {
		syscallarg(int) fd;
		syscallarg(int) flags;
	} */ *uap;
{
	struct file *fp;
	int error;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	error = setfflags(td, (struct vnode *) fp->f_data, SCARG(uap, flags));
	fdrop(fp, td);
	return (error);
}

/*
 * Common implementation code for chmod(), lchmod() and fchmod().
 */
static int
setfmode(td, vp, mode)
	struct thread *td;
	struct vnode *vp;
	int mode;
{
	int error;
	struct mount *mp;
	struct vattr vattr;

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	VATTR_NULL(&vattr);
	vattr.va_mode = mode & ALLPERMS;
	error = VOP_SETATTR(vp, &vattr, td->td_ucred, td);
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return error;
}

/*
 * Change mode of a file given path name.
 */
#ifndef _SYS_SYSPROTO_H_
struct chmod_args {
	char	*path;
	int	mode;
};
#endif
/* ARGSUSED */
int
chmod(td, uap)
	struct thread *td;
	register struct chmod_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setfmode(td, nd.ni_vp, SCARG(uap, mode));
	vrele(nd.ni_vp);
	return error;
}

/*
 * Change mode of a file given path name (don't follow links.)
 */
#ifndef _SYS_SYSPROTO_H_
struct lchmod_args {
	char	*path;
	int	mode;
};
#endif
/* ARGSUSED */
int
lchmod(td, uap)
	struct thread *td;
	register struct lchmod_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setfmode(td, nd.ni_vp, SCARG(uap, mode));
	vrele(nd.ni_vp);
	return error;
}

/*
 * Change mode of a file given a file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct fchmod_args {
	int	fd;
	int	mode;
};
#endif
/* ARGSUSED */
int
fchmod(td, uap)
	struct thread *td;
	register struct fchmod_args /* {
		syscallarg(int) fd;
		syscallarg(int) mode;
	} */ *uap;
{
	struct file *fp;
	struct vnode *vp;
	int error;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	vp = (struct vnode *)fp->f_data;
	error = setfmode(td, (struct vnode *)fp->f_data, SCARG(uap, mode));
	fdrop(fp, td);
	return (error);
}

/*
 * Common implementation for chown(), lchown(), and fchown()
 */
static int
setfown(td, vp, uid, gid)
	struct thread *td;
	struct vnode *vp;
	uid_t uid;
	gid_t gid;
{
	int error;
	struct mount *mp;
	struct vattr vattr;

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	VATTR_NULL(&vattr);
	vattr.va_uid = uid;
	vattr.va_gid = gid;
	error = VOP_SETATTR(vp, &vattr, td->td_ucred, td);
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return error;
}

/*
 * Set ownership given a path name.
 */
#ifndef _SYS_SYSPROTO_H_
struct chown_args {
	char	*path;
	int	uid;
	int	gid;
};
#endif
/* ARGSUSED */
int
chown(td, uap)
	struct thread *td;
	register struct chown_args /* {
		syscallarg(char *) path;
		syscallarg(int) uid;
		syscallarg(int) gid;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setfown(td, nd.ni_vp, SCARG(uap, uid), SCARG(uap, gid));
	vrele(nd.ni_vp);
	return (error);
}

/*
 * Set ownership given a path name, do not cross symlinks.
 */
#ifndef _SYS_SYSPROTO_H_
struct lchown_args {
	char	*path;
	int	uid;
	int	gid;
};
#endif
/* ARGSUSED */
int
lchown(td, uap)
	struct thread *td;
	register struct lchown_args /* {
		syscallarg(char *) path;
		syscallarg(int) uid;
		syscallarg(int) gid;
	} */ *uap;
{
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, NOFOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setfown(td, nd.ni_vp, SCARG(uap, uid), SCARG(uap, gid));
	vrele(nd.ni_vp);
	return (error);
}

/*
 * Set ownership given a file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct fchown_args {
	int	fd;
	int	uid;
	int	gid;
};
#endif
/* ARGSUSED */
int
fchown(td, uap)
	struct thread *td;
	register struct fchown_args /* {
		syscallarg(int) fd;
		syscallarg(int) uid;
		syscallarg(int) gid;
	} */ *uap;
{
	struct file *fp;
	struct vnode *vp;
	int error;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	vp = (struct vnode *)fp->f_data;
	error = setfown(td, (struct vnode *)fp->f_data,
		SCARG(uap, uid), SCARG(uap, gid));
	fdrop(fp, td);
	return (error);
}

/*
 * Common implementation code for utimes(), lutimes(), and futimes().
 */
static int
getutimes(usrtvp, tsp)
	const struct timeval *usrtvp;
	struct timespec *tsp;
{
	struct timeval tv[2];
	int error;

	if (usrtvp == NULL) {
		microtime(&tv[0]);
		TIMEVAL_TO_TIMESPEC(&tv[0], &tsp[0]);
		tsp[1] = tsp[0];
	} else {
		if ((error = copyin(usrtvp, tv, sizeof (tv))) != 0)
			return (error);
		TIMEVAL_TO_TIMESPEC(&tv[0], &tsp[0]);
		TIMEVAL_TO_TIMESPEC(&tv[1], &tsp[1]);
	}
	return 0;
}

/*
 * Common implementation code for utimes(), lutimes(), and futimes().
 */
static int
setutimes(td, vp, ts, nullflag)
	struct thread *td;
	struct vnode *vp;
	const struct timespec *ts;
	int nullflag;
{
	int error;
	struct mount *mp;
	struct vattr vattr;

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	VATTR_NULL(&vattr);
	vattr.va_atime = ts[0];
	vattr.va_mtime = ts[1];
	if (nullflag)
		vattr.va_vaflags |= VA_UTIMES_NULL;
	error = VOP_SETATTR(vp, &vattr, td->td_ucred, td);
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return error;
}

/*
 * Set the access and modification times of a file.
 */
#ifndef _SYS_SYSPROTO_H_
struct utimes_args {
	char	*path;
	struct	timeval *tptr;
};
#endif
/* ARGSUSED */
int
utimes(td, uap)
	struct thread *td;
	register struct utimes_args /* {
		syscallarg(char *) path;
		syscallarg(struct timeval *) tptr;
	} */ *uap;
{
	struct timespec ts[2];
	struct timeval *usrtvp;
	int error;
	struct nameidata nd;

	usrtvp = SCARG(uap, tptr);
	if ((error = getutimes(usrtvp, ts)) != 0)
		return (error);
	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setutimes(td, nd.ni_vp, ts, usrtvp == NULL);
	vrele(nd.ni_vp);
	return (error);
}

/*
 * Set the access and modification times of a file.
 */
#ifndef _SYS_SYSPROTO_H_
struct lutimes_args {
	char	*path;
	struct	timeval *tptr;
};
#endif
/* ARGSUSED */
int
lutimes(td, uap)
	struct thread *td;
	register struct lutimes_args /* {
		syscallarg(char *) path;
		syscallarg(struct timeval *) tptr;
	} */ *uap;
{
	struct timespec ts[2];
	struct timeval *usrtvp;
	int error;
	struct nameidata nd;

	usrtvp = SCARG(uap, tptr);
	if ((error = getutimes(usrtvp, ts)) != 0)
		return (error);
	NDINIT(&nd, LOOKUP, NOFOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	error = setutimes(td, nd.ni_vp, ts, usrtvp == NULL);
	vrele(nd.ni_vp);
	return (error);
}

/*
 * Set the access and modification times of a file.
 */
#ifndef _SYS_SYSPROTO_H_
struct futimes_args {
	int	fd;
	struct	timeval *tptr;
};
#endif
/* ARGSUSED */
int
futimes(td, uap)
	struct thread *td;
	register struct futimes_args /* {
		syscallarg(int ) fd;
		syscallarg(struct timeval *) tptr;
	} */ *uap;
{
	struct timespec ts[2];
	struct file *fp;
	struct timeval *usrtvp;
	int error;

	usrtvp = SCARG(uap, tptr);
	if ((error = getutimes(usrtvp, ts)) != 0)
		return (error);
	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	error = setutimes(td, (struct vnode *)fp->f_data, ts, usrtvp == NULL);
	fdrop(fp, td);
	return (error);
}

/*
 * Truncate a file given its path name.
 */
#ifndef _SYS_SYSPROTO_H_
struct truncate_args {
	char	*path;
	int	pad;
	off_t	length;
};
#endif
/* ARGSUSED */
int
truncate(td, uap)
	struct thread *td;
	register struct truncate_args /* {
		syscallarg(char *) path;
		syscallarg(int) pad;
		syscallarg(off_t) length;
	} */ *uap;
{
	struct mount *mp;
	struct vnode *vp;
	struct vattr vattr;
	int error;
	struct nameidata nd;

	if (uap->length < 0)
		return(EINVAL);
	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) {
		vrele(vp);
		return (error);
	}
	NDFREE(&nd, NDF_ONLY_PNBUF);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	if (vp->v_type == VDIR)
		error = EISDIR;
	else if ((error = vn_writechk(vp)) == 0 &&
	    (error = VOP_ACCESS(vp, VWRITE, td->td_ucred, td)) == 0) {
		VATTR_NULL(&vattr);
		vattr.va_size = SCARG(uap, length);
		error = VOP_SETATTR(vp, &vattr, td->td_ucred, td);
	}
	vput(vp);
	vn_finished_write(mp);
	return (error);
}

/*
 * Truncate a file given a file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct ftruncate_args {
	int	fd;
	int	pad;
	off_t	length;
};
#endif
/* ARGSUSED */
int
ftruncate(td, uap)
	struct thread *td;
	register struct ftruncate_args /* {
		syscallarg(int) fd;
		syscallarg(int) pad;
		syscallarg(off_t) length;
	} */ *uap;
{
	struct mount *mp;
	struct vattr vattr;
	struct vnode *vp;
	struct file *fp;
	int error;

	if (uap->length < 0)
		return(EINVAL);
	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	if ((fp->f_flag & FWRITE) == 0) {
		fdrop(fp, td);
		return (EINVAL);
	}
	vp = (struct vnode *)fp->f_data;
	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) {
		fdrop(fp, td);
		return (error);
	}
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	if (vp->v_type == VDIR)
		error = EISDIR;
	else if ((error = vn_writechk(vp)) == 0) {
		VATTR_NULL(&vattr);
		vattr.va_size = SCARG(uap, length);
		error = VOP_SETATTR(vp, &vattr, fp->f_cred, td);
	}
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	fdrop(fp, td);
	return (error);
}

#if defined(COMPAT_43) || defined(COMPAT_SUNOS)
/*
 * Truncate a file given its path name.
 */
#ifndef _SYS_SYSPROTO_H_
struct otruncate_args {
	char	*path;
	long	length;
};
#endif
/* ARGSUSED */
int
otruncate(td, uap)
	struct thread *td;
	register struct otruncate_args /* {
		syscallarg(char *) path;
		syscallarg(long) length;
	} */ *uap;
{
	struct truncate_args /* {
		syscallarg(char *) path;
		syscallarg(int) pad;
		syscallarg(off_t) length;
	} */ nuap;

	SCARG(&nuap, path) = SCARG(uap, path);
	SCARG(&nuap, length) = SCARG(uap, length);
	return (truncate(td, &nuap));
}

/*
 * Truncate a file given a file descriptor.
 */
#ifndef _SYS_SYSPROTO_H_
struct oftruncate_args {
	int	fd;
	long	length;
};
#endif
/* ARGSUSED */
int
oftruncate(td, uap)
	struct thread *td;
	register struct oftruncate_args /* {
		syscallarg(int) fd;
		syscallarg(long) length;
	} */ *uap;
{
	struct ftruncate_args /* {
		syscallarg(int) fd;
		syscallarg(int) pad;
		syscallarg(off_t) length;
	} */ nuap;

	SCARG(&nuap, fd) = SCARG(uap, fd);
	SCARG(&nuap, length) = SCARG(uap, length);
	return (ftruncate(td, &nuap));
}
#endif /* COMPAT_43 || COMPAT_SUNOS */

/*
 * Sync an open file.
 */
#ifndef _SYS_SYSPROTO_H_
struct fsync_args {
	int	fd;
};
#endif
/* ARGSUSED */
int
fsync(td, uap)
	struct thread *td;
	struct fsync_args /* {
		syscallarg(int) fd;
	} */ *uap;
{
	struct vnode *vp;
	struct mount *mp;
	struct file *fp;
	vm_object_t obj;
	int error;

	GIANT_REQUIRED;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	vp = (struct vnode *)fp->f_data;
	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) {
		fdrop(fp, td);
		return (error);
	}
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	if (VOP_GETVOBJECT(vp, &obj) == 0) {
		vm_object_page_clean(obj, 0, 0, 0);
	}
	error = VOP_FSYNC(vp, fp->f_cred, MNT_WAIT, td);
#ifdef SOFTUPDATES
	if (error == 0 && vp->v_mount && (vp->v_mount->mnt_flag & MNT_SOFTDEP))
	    error = softdep_fsync(vp);
#endif

	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	fdrop(fp, td);
	return (error);
}

/*
 * Rename files.  Source and destination must either both be directories,
 * or both not be directories.  If target is a directory, it must be empty.
 */
#ifndef _SYS_SYSPROTO_H_
struct rename_args {
	char	*from;
	char	*to;
};
#endif
/* ARGSUSED */
int
rename(td, uap)
	struct thread *td;
	register struct rename_args /* {
		syscallarg(char *) from;
		syscallarg(char *) to;
	} */ *uap;
{
	struct mount *mp;
	struct vnode *tvp, *fvp, *tdvp;
	struct nameidata fromnd, tond;
	int error;

	bwillwrite();
	NDINIT(&fromnd, DELETE, WANTPARENT | SAVESTART, UIO_USERSPACE,
	    SCARG(uap, from), td);
	if ((error = namei(&fromnd)) != 0)
		return (error);
	fvp = fromnd.ni_vp;
	if ((error = vn_start_write(fvp, &mp, V_WAIT | PCATCH)) != 0) {
		NDFREE(&fromnd, NDF_ONLY_PNBUF);
		vrele(fromnd.ni_dvp);
		vrele(fvp);
		goto out1;
	}
	NDINIT(&tond, RENAME, LOCKPARENT | LOCKLEAF | NOCACHE | SAVESTART | NOOBJ,
	    UIO_USERSPACE, SCARG(uap, to), td);
	if (fromnd.ni_vp->v_type == VDIR)
		tond.ni_cnd.cn_flags |= WILLBEDIR;
	if ((error = namei(&tond)) != 0) {
		/* Translate error code for rename("dir1", "dir2/."). */
		if (error == EISDIR && fvp->v_type == VDIR)
			error = EINVAL;
		NDFREE(&fromnd, NDF_ONLY_PNBUF);
		vrele(fromnd.ni_dvp);
		vrele(fvp);
		goto out1;
	}
	tdvp = tond.ni_dvp;
	tvp = tond.ni_vp;
	if (tvp != NULL) {
		if (fvp->v_type == VDIR && tvp->v_type != VDIR) {
			error = ENOTDIR;
			goto out;
		} else if (fvp->v_type != VDIR && tvp->v_type == VDIR) {
			error = EISDIR;
			goto out;
		}
	}
	if (fvp == tdvp)
		error = EINVAL;
	/*
	 * If source is the same as the destination (that is the
	 * same inode number with the same name in the same directory),
	 * then there is nothing to do.
	 */
	if (fvp == tvp && fromnd.ni_dvp == tdvp &&
	    fromnd.ni_cnd.cn_namelen == tond.ni_cnd.cn_namelen &&
	    !bcmp(fromnd.ni_cnd.cn_nameptr, tond.ni_cnd.cn_nameptr,
	      fromnd.ni_cnd.cn_namelen))
		error = -1;
out:
	if (!error) {
		VOP_LEASE(tdvp, td, td->td_ucred, LEASE_WRITE);
		if (fromnd.ni_dvp != tdvp) {
			VOP_LEASE(fromnd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
		}
		if (tvp) {
			VOP_LEASE(tvp, td, td->td_ucred, LEASE_WRITE);
		}
		error = VOP_RENAME(fromnd.ni_dvp, fromnd.ni_vp, &fromnd.ni_cnd,
				   tond.ni_dvp, tond.ni_vp, &tond.ni_cnd);
		NDFREE(&fromnd, NDF_ONLY_PNBUF);
		NDFREE(&tond, NDF_ONLY_PNBUF);
	} else {
		NDFREE(&fromnd, NDF_ONLY_PNBUF);
		NDFREE(&tond, NDF_ONLY_PNBUF);
		if (tdvp == tvp)
			vrele(tdvp);
		else
			vput(tdvp);
		if (tvp)
			vput(tvp);
		vrele(fromnd.ni_dvp);
		vrele(fvp);
	}
	vrele(tond.ni_startdir);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(fromnd.ni_dvp, "rename");
	ASSERT_VOP_UNLOCKED(fromnd.ni_vp, "rename");
	ASSERT_VOP_UNLOCKED(tond.ni_dvp, "rename");
	ASSERT_VOP_UNLOCKED(tond.ni_vp, "rename");
out1:
	if (fromnd.ni_startdir)
		vrele(fromnd.ni_startdir);
	if (error == -1)
		return (0);
	return (error);
}

/*
 * Make a directory file.
 */
#ifndef _SYS_SYSPROTO_H_
struct mkdir_args {
	char	*path;
	int	mode;
};
#endif
/* ARGSUSED */
int
mkdir(td, uap)
	struct thread *td;
	register struct mkdir_args /* {
		syscallarg(char *) path;
		syscallarg(int) mode;
	} */ *uap;
{

	return vn_mkdir(uap->path, uap->mode, UIO_USERSPACE, td);
}

int
vn_mkdir(path, mode, segflg, td)
	char *path;
	int mode;
	enum uio_seg segflg;
	struct thread *td;
{
	struct mount *mp;
	struct vnode *vp;
	struct vattr vattr;
	int error;
	struct nameidata nd;

restart:
	bwillwrite();
	NDINIT(&nd, CREATE, LOCKPARENT, segflg, path, td);
	nd.ni_cnd.cn_flags |= WILLBEDIR;
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	if (vp != NULL) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vrele(vp);
		vput(nd.ni_dvp);
		return (EEXIST);
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		vput(nd.ni_dvp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VATTR_NULL(&vattr);
	vattr.va_type = VDIR;
	FILEDESC_LOCK(td->td_proc->p_fd);
	vattr.va_mode = (mode & ACCESSPERMS) &~ td->td_proc->p_fd->fd_cmask;
	FILEDESC_UNLOCK(td->td_proc->p_fd);
	VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
	error = VOP_MKDIR(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vput(nd.ni_dvp);
	if (!error)
		vput(nd.ni_vp);
	vn_finished_write(mp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "mkdir");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "mkdir");
	return (error);
}

/*
 * Remove a directory file.
 */
#ifndef _SYS_SYSPROTO_H_
struct rmdir_args {
	char	*path;
};
#endif
/* ARGSUSED */
int
rmdir(td, uap)
	struct thread *td;
	struct rmdir_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	struct mount *mp;
	struct vnode *vp;
	int error;
	struct nameidata nd;

restart:
	bwillwrite();
	NDINIT(&nd, DELETE, LOCKPARENT | LOCKLEAF, UIO_USERSPACE,
	    SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	if (vp->v_type != VDIR) {
		error = ENOTDIR;
		goto out;
	}
	/*
	 * No rmdir "." please.
	 */
	if (nd.ni_dvp == vp) {
		error = EINVAL;
		goto out;
	}
	/*
	 * The root of a mounted filesystem cannot be deleted.
	 */
	if (vp->v_flag & VROOT) {
		error = EBUSY;
		goto out;
	}
	if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) {
		NDFREE(&nd, NDF_ONLY_PNBUF);
		if (nd.ni_dvp == vp)
			vrele(nd.ni_dvp);
		else
			vput(nd.ni_dvp);
		vput(vp);
		if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0)
			return (error);
		goto restart;
	}
	VOP_LEASE(nd.ni_dvp, td, td->td_ucred, LEASE_WRITE);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	error = VOP_RMDIR(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd);
	vn_finished_write(mp);
out:
	NDFREE(&nd, NDF_ONLY_PNBUF);
	if (nd.ni_dvp == vp)
		vrele(nd.ni_dvp);
	else
		vput(nd.ni_dvp);
	vput(vp);
	ASSERT_VOP_UNLOCKED(nd.ni_dvp, "rmdir");
	ASSERT_VOP_UNLOCKED(nd.ni_vp, "rmdir");
	return (error);
}

#ifdef COMPAT_43
/*
 * Read a block of directory entries in a file system independent format.
 */
#ifndef _SYS_SYSPROTO_H_
struct ogetdirentries_args {
	int	fd;
	char	*buf;
	u_int	count;
	long	*basep;
};
#endif
int
ogetdirentries(td, uap)
	struct thread *td;
	register struct ogetdirentries_args /* {
		syscallarg(int) fd;
		syscallarg(char *) buf;
		syscallarg(u_int) count;
		syscallarg(long *) basep;
	} */ *uap;
{
	struct vnode *vp;
	struct file *fp;
	struct uio auio, kuio;
	struct iovec aiov, kiov;
	struct dirent *dp, *edp;
	caddr_t dirbuf;
	int error, eofflag, readcnt;
	long loff;

	/* XXX arbitrary sanity limit on `count'. */
	if (SCARG(uap, count) > 64 * 1024)
		return (EINVAL);
	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	if ((fp->f_flag & FREAD) == 0) {
		fdrop(fp, td);
		return (EBADF);
	}
	vp = (struct vnode *)fp->f_data;
unionread:
	if (vp->v_type != VDIR) {
		fdrop(fp, td);
		return (EINVAL);
	}
	aiov.iov_base = SCARG(uap, buf);
	aiov.iov_len = SCARG(uap, count);
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	auio.uio_rw = UIO_READ;
	auio.uio_segflg = UIO_USERSPACE;
	auio.uio_td = td;
	auio.uio_resid = SCARG(uap, count);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	loff = auio.uio_offset = fp->f_offset;
#	if (BYTE_ORDER != LITTLE_ENDIAN)
		if (vp->v_mount->mnt_maxsymlinklen <= 0) {
			error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag,
			    NULL, NULL);
			fp->f_offset = auio.uio_offset;
		} else
#	endif
	{
		kuio = auio;
		kuio.uio_iov = &kiov;
		kuio.uio_segflg = UIO_SYSSPACE;
		kiov.iov_len = SCARG(uap, count);
		MALLOC(dirbuf, caddr_t, SCARG(uap, count), M_TEMP, M_WAITOK);
		kiov.iov_base = dirbuf;
		error = VOP_READDIR(vp, &kuio, fp->f_cred, &eofflag,
			    NULL, NULL);
		fp->f_offset = kuio.uio_offset;
		if (error == 0) {
			readcnt = SCARG(uap, count) - kuio.uio_resid;
			edp = (struct dirent *)&dirbuf[readcnt];
			for (dp = (struct dirent *)dirbuf; dp < edp; ) {
#				if (BYTE_ORDER == LITTLE_ENDIAN)
					/*
					 * The expected low byte of
					 * dp->d_namlen is our dp->d_type.
					 * The high MBZ byte of dp->d_namlen
					 * is our dp->d_namlen.
					 */
					dp->d_type = dp->d_namlen;
					dp->d_namlen = 0;
#				else
					/*
					 * The dp->d_type is the high byte
					 * of the expected dp->d_namlen,
					 * so must be zero'ed.
					 */
					dp->d_type = 0;
#				endif
				if (dp->d_reclen > 0) {
					dp = (struct dirent *)
					    ((char *)dp + dp->d_reclen);
				} else {
					error = EIO;
					break;
				}
			}
			if (dp >= edp)
				error = uiomove(dirbuf, readcnt, &auio);
		}
		FREE(dirbuf, M_TEMP);
	}
	VOP_UNLOCK(vp, 0, td);
	if (error) {
		fdrop(fp, td);
		return (error);
	}
	if (SCARG(uap, count) == auio.uio_resid) {
		if (union_dircheckp) {
			error = union_dircheckp(td, &vp, fp);
			if (error == -1)
				goto unionread;
			if (error) {
				fdrop(fp, td);
				return (error);
			}
		}
		if ((vp->v_flag & VROOT) &&
		    (vp->v_mount->mnt_flag & MNT_UNION)) {
			struct vnode *tvp = vp;
			vp = vp->v_mount->mnt_vnodecovered;
			VREF(vp);
			fp->f_data = (caddr_t) vp;
			fp->f_offset = 0;
			vrele(tvp);
			goto unionread;
		}
	}
	error = copyout((caddr_t)&loff, (caddr_t)SCARG(uap, basep),
	    sizeof(long));
	fdrop(fp, td);
	td->td_retval[0] = SCARG(uap, count) - auio.uio_resid;
	return (error);
}
#endif /* COMPAT_43 */

/*
 * Read a block of directory entries in a file system independent format.
 */
#ifndef _SYS_SYSPROTO_H_
struct getdirentries_args {
	int	fd;
	char	*buf;
	u_int	count;
	long	*basep;
};
#endif
int
getdirentries(td, uap)
	struct thread *td;
	register struct getdirentries_args /* {
		syscallarg(int) fd;
		syscallarg(char *) buf;
		syscallarg(u_int) count;
		syscallarg(long *) basep;
	} */ *uap;
{
	struct vnode *vp;
	struct file *fp;
	struct uio auio;
	struct iovec aiov;
	long loff;
	int error, eofflag;

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	if ((fp->f_flag & FREAD) == 0) {
		fdrop(fp, td);
		return (EBADF);
	}
	vp = (struct vnode *)fp->f_data;
unionread:
	if (vp->v_type != VDIR) {
		fdrop(fp, td);
		return (EINVAL);
	}
	aiov.iov_base = SCARG(uap, buf);
	aiov.iov_len = SCARG(uap, count);
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	auio.uio_rw = UIO_READ;
	auio.uio_segflg = UIO_USERSPACE;
	auio.uio_td = td;
	auio.uio_resid = SCARG(uap, count);
	/* vn_lock(vp, LK_SHARED | LK_RETRY, td); */
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
	loff = auio.uio_offset = fp->f_offset;
	error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, NULL, NULL);
	fp->f_offset = auio.uio_offset;
	VOP_UNLOCK(vp, 0, td);
	if (error) {
		fdrop(fp, td);
		return (error);
	}
	if (SCARG(uap, count) == auio.uio_resid) {
		if (union_dircheckp) {
			error = union_dircheckp(td, &vp, fp);
			if (error == -1)
				goto unionread;
			if (error) {
				fdrop(fp, td);
				return (error);
			}
		}
		if ((vp->v_flag & VROOT) &&
		    (vp->v_mount->mnt_flag & MNT_UNION)) {
			struct vnode *tvp = vp;
			vp = vp->v_mount->mnt_vnodecovered;
			VREF(vp);
			fp->f_data = (caddr_t) vp;
			fp->f_offset = 0;
			vrele(tvp);
			goto unionread;
		}
	}
	if (SCARG(uap, basep) != NULL) {
		error = copyout((caddr_t)&loff, (caddr_t)SCARG(uap, basep),
		    sizeof(long));
	}
	td->td_retval[0] = SCARG(uap, count) - auio.uio_resid;
	fdrop(fp, td);
	return (error);
}
#ifndef _SYS_SYSPROTO_H_
struct getdents_args {
	int fd;
	char *buf;
	size_t count;
};
#endif
int
getdents(td, uap)
	struct thread *td;
	register struct getdents_args /* {
		syscallarg(int) fd;
		syscallarg(char *) buf;
		syscallarg(u_int) count;
	} */ *uap;
{
	struct getdirentries_args ap;
	ap.fd = uap->fd;
	ap.buf = uap->buf;
	ap.count = uap->count;
	ap.basep = NULL;
	return getdirentries(td, &ap);
}

/*
 * Set the mode mask for creation of filesystem nodes.
 *
 * MP SAFE
 */
#ifndef _SYS_SYSPROTO_H_
struct umask_args {
	int	newmask;
};
#endif
int
umask(td, uap)
	struct thread *td;
	struct umask_args /* {
		syscallarg(int) newmask;
	} */ *uap;
{
	register struct filedesc *fdp;

	FILEDESC_LOCK(td->td_proc->p_fd);
	fdp = td->td_proc->p_fd;
	td->td_retval[0] = fdp->fd_cmask;
	fdp->fd_cmask = SCARG(uap, newmask) & ALLPERMS;
	FILEDESC_UNLOCK(td->td_proc->p_fd);
	return (0);
}

/*
 * Void all references to file by ripping underlying filesystem
 * away from vnode.
 */
#ifndef _SYS_SYSPROTO_H_
struct revoke_args {
	char	*path;
};
#endif
/* ARGSUSED */
int
revoke(td, uap)
	struct thread *td;
	register struct revoke_args /* {
		syscallarg(char *) path;
	} */ *uap;
{
	struct mount *mp;
	struct vnode *vp;
	struct vattr vattr;
	int error;
	struct nameidata nd;

	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE, SCARG(uap, path),
	    td);
	if ((error = namei(&nd)) != 0)
		return (error);
	vp = nd.ni_vp;
	NDFREE(&nd, NDF_ONLY_PNBUF);
	if (vp->v_type != VCHR) {
		vput(vp);
		return (EINVAL);
	}
	error = VOP_GETATTR(vp, &vattr, td->td_ucred, td);
	if (error) {
		vput(vp);
		return (error);
	}
	VOP_UNLOCK(vp, 0, td);
	if (td->td_ucred->cr_uid != vattr.va_uid) {
		error = suser_xxx(0, td->td_proc, PRISON_ROOT);
		if (error)
			goto out;
	}
	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		goto out;
	if (vcount(vp) > 1)
		VOP_REVOKE(vp, REVOKEALL);
	vn_finished_write(mp);
out:
	vrele(vp);
	return (error);
}

/*
 * Convert a user file descriptor to a kernel file entry.
 * The file entry is locked upon returning.
 */
int
getvnode(fdp, fd, fpp)
	struct filedesc *fdp;
	int fd;
	struct file **fpp;
{
	int error;
	struct file *fp;

	fp = NULL;
	if (fdp == NULL)
		error = EBADF;
	else {
		FILEDESC_LOCK(fdp);
		if ((u_int)fd >= fdp->fd_nfiles ||
		    (fp = fdp->fd_ofiles[fd]) == NULL)
			error = EBADF;
		else if (fp->f_type != DTYPE_VNODE && fp->f_type != DTYPE_FIFO) {
			fp = NULL;
			error = EINVAL;
		} else {
			fhold(fp);
			error = 0;
		}
		FILEDESC_UNLOCK(fdp);
	}
	*fpp = fp;
	return (error);
}
/*
 * Get (NFS) file handle
 */
#ifndef _SYS_SYSPROTO_H_
struct getfh_args {
	char	*fname;
	fhandle_t *fhp;
};
#endif
int
getfh(td, uap)
	struct thread *td;
	register struct getfh_args *uap;
{
	struct nameidata nd;
	fhandle_t fh;
	register struct vnode *vp;
	int error;

	/*
	 * Must be super user
	 */
	error = suser_td(td);
	if (error)
		return (error);
	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE, uap->fname, td);
	error = namei(&nd);
	if (error)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);
	vp = nd.ni_vp;
	bzero(&fh, sizeof(fh));
	fh.fh_fsid = vp->v_mount->mnt_stat.f_fsid;
	error = VFS_VPTOFH(vp, &fh.fh_fid);
	vput(vp);
	if (error)
		return (error);
	error = copyout(&fh, uap->fhp, sizeof (fh));
	return (error);
}

/*
 * syscall for the rpc.lockd to use to translate a NFS file handle into
 * an open descriptor.
 *
 * warning: do not remove the suser() call or this becomes one giant
 * security hole.
 */
#ifndef _SYS_SYSPROTO_H_
struct fhopen_args {
	const struct fhandle *u_fhp;
	int flags;
};
#endif
int
fhopen(td, uap)
	struct thread *td;
	struct fhopen_args /* {
		syscallarg(const struct fhandle *) u_fhp;
		syscallarg(int) flags;
	} */ *uap;
{
	struct proc *p = td->td_proc;
	struct mount *mp;
	struct vnode *vp;
	struct fhandle fhp;
	struct vattr vat;
	struct vattr *vap = &vat;
	struct flock lf;
	struct file *fp;
	register struct filedesc *fdp = p->p_fd;
	int fmode, mode, error, type;
	struct file *nfp; 
	int indx;

	/*
	 * Must be super user
	 */
	error = suser_td(td);
	if (error)
		return (error);

	fmode = FFLAGS(SCARG(uap, flags));
	/* why not allow a non-read/write open for our lockd? */
	if (((fmode & (FREAD | FWRITE)) == 0) || (fmode & O_CREAT))
		return (EINVAL);
	error = copyin(SCARG(uap,u_fhp), &fhp, sizeof(fhp));
	if (error)
		return(error);
	/* find the mount point */
	mp = vfs_getvfs(&fhp.fh_fsid);
	if (mp == NULL)
		return (ESTALE);
	/* now give me my vnode, it gets returned to me locked */
	error = VFS_FHTOVP(mp, &fhp.fh_fid, &vp);
	if (error)
		return (error);
 	/*
	 * from now on we have to make sure not
	 * to forget about the vnode
	 * any error that causes an abort must vput(vp) 
	 * just set error = err and 'goto bad;'.
	 */

	/* 
	 * from vn_open 
	 */
	if (vp->v_type == VLNK) {
		error = EMLINK;
		goto bad;
	}
	if (vp->v_type == VSOCK) {
		error = EOPNOTSUPP;
		goto bad;
	}
	mode = 0;
	if (fmode & (FWRITE | O_TRUNC)) {
		if (vp->v_type == VDIR) {
			error = EISDIR;
			goto bad;
		}
		error = vn_writechk(vp);
		if (error)
			goto bad;
		mode |= VWRITE;
	}
	if (fmode & FREAD)
		mode |= VREAD;
	if (mode) {
		error = VOP_ACCESS(vp, mode, td->td_ucred, td);
		if (error)
			goto bad;
	}
	if (fmode & O_TRUNC) {
		VOP_UNLOCK(vp, 0, td);				/* XXX */
		if ((error = vn_start_write(NULL, &mp, V_WAIT | PCATCH)) != 0) {
			vrele(vp);
			return (error);
		}
		VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);	/* XXX */
		VATTR_NULL(vap);
		vap->va_size = 0;
		error = VOP_SETATTR(vp, vap, td->td_ucred, td);
		vn_finished_write(mp);
		if (error)
			goto bad;
	}
	error = VOP_OPEN(vp, fmode, td->td_ucred, td);
	if (error)
		goto bad;
	/*
	 * Make sure that a VM object is created for VMIO support.
	 */
	if (vn_canvmio(vp) == TRUE) {
		if ((error = vfs_object_create(vp, td, td->td_ucred)) != 0)
			goto bad;
	}
	if (fmode & FWRITE)
		vp->v_writecount++;

	/*
	 * end of vn_open code 
	 */

	if ((error = falloc(td, &nfp, &indx)) != 0) {
		if (fmode & FWRITE)
			vp->v_writecount--;
		goto bad;
	}
	fp = nfp;	

	/*
	 * Hold an extra reference to avoid having fp ripped out 
	 * from under us while we block in the lock op
	 */
	fhold(fp);
	nfp->f_data = (caddr_t)vp;
	nfp->f_flag = fmode & FMASK;
	nfp->f_ops = &vnops;
	nfp->f_type = DTYPE_VNODE;
	if (fmode & (O_EXLOCK | O_SHLOCK)) {
		lf.l_whence = SEEK_SET;
		lf.l_start = 0;
		lf.l_len = 0;
		if (fmode & O_EXLOCK)
			lf.l_type = F_WRLCK;
		else
			lf.l_type = F_RDLCK;
		type = F_FLOCK;
		if ((fmode & FNONBLOCK) == 0)
			type |= F_WAIT;
		VOP_UNLOCK(vp, 0, td);
		if ((error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type)) != 0) {
			/*
			 * The lock request failed.  Normally close the
			 * descriptor but handle the case where someone might
			 * have dup()d or close()d it when we weren't looking.
			 */
			FILEDESC_LOCK(fdp);
			if (fdp->fd_ofiles[indx] == fp) {
				fdp->fd_ofiles[indx] = NULL;
				FILEDESC_UNLOCK(fdp);
				fdrop(fp, td);
			} else
				FILEDESC_UNLOCK(fdp);
			/*
			 * release our private reference
			 */
			fdrop(fp, td);
			return(error);
		}
		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
		fp->f_flag |= FHASLOCK;
	}
	if ((vp->v_type == VREG) && (VOP_GETVOBJECT(vp, NULL) != 0))
		vfs_object_create(vp, td, td->td_ucred);

	VOP_UNLOCK(vp, 0, td);
	fdrop(fp, td);
	td->td_retval[0] = indx;
	return (0);

bad:
	vput(vp);
	return (error);
}

/*
 * Stat an (NFS) file handle.
 */
#ifndef _SYS_SYSPROTO_H_
struct fhstat_args {
	struct fhandle *u_fhp;
	struct stat *sb;
};
#endif
int
fhstat(td, uap)
	struct thread *td;
	register struct fhstat_args /* {
		syscallarg(struct fhandle *) u_fhp;
		syscallarg(struct stat *) sb;
	} */ *uap;
{
	struct stat sb;
	fhandle_t fh;
	struct mount *mp;
	struct vnode *vp;
	int error;

	/*
	 * Must be super user
	 */
	error = suser_td(td);
	if (error)
		return (error);
	
	error = copyin(SCARG(uap, u_fhp), &fh, sizeof(fhandle_t));
	if (error)
		return (error);

	if ((mp = vfs_getvfs(&fh.fh_fsid)) == NULL)
		return (ESTALE);
	if ((error = VFS_FHTOVP(mp, &fh.fh_fid, &vp)))
		return (error);
	error = vn_stat(vp, &sb, td);
	vput(vp);
	if (error)
		return (error);
	error = copyout(&sb, SCARG(uap, sb), sizeof(sb));
	return (error);
}

/*
 * Implement fstatfs() for (NFS) file handles.
 */
#ifndef _SYS_SYSPROTO_H_
struct fhstatfs_args {
	struct fhandle *u_fhp;
	struct statfs *buf;
};
#endif
int
fhstatfs(td, uap)
	struct thread *td;
	struct fhstatfs_args /* {
		syscallarg(struct fhandle) *u_fhp;
		syscallarg(struct statfs) *buf;
	} */ *uap;
{
	struct statfs *sp;
	struct mount *mp;
	struct vnode *vp;
	struct statfs sb;
	fhandle_t fh;
	int error;

	/*
	 * Must be super user
	 */
	error = suser_td(td);
	if (error)
		return (error);

	if ((error = copyin(SCARG(uap, u_fhp), &fh, sizeof(fhandle_t))) != 0)
		return (error);

	if ((mp = vfs_getvfs(&fh.fh_fsid)) == NULL)
		return (ESTALE);
	if ((error = VFS_FHTOVP(mp, &fh.fh_fid, &vp)))
		return (error);
	mp = vp->v_mount;
	sp = &mp->mnt_stat;
	vput(vp);
	if ((error = VFS_STATFS(mp, sp, td)) != 0)
		return (error);
	sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
	if (suser_xxx(td->td_ucred, 0, 0)) {
		bcopy((caddr_t)sp, (caddr_t)&sb, sizeof(sb));
		sb.f_fsid.val[0] = sb.f_fsid.val[1] = 0;
		sp = &sb;
	}
	return (copyout(sp, SCARG(uap, buf), sizeof(*sp)));
}

/*
 * Syscall to push extended attribute configuration information into the
 * VFS.  Accepts a path, which it converts to a mountpoint, as well as
 * a command (int cmd), and attribute name and misc data.  For now, the
 * attribute name is left in userspace for consumption by the VFS_op.
 * It will probably be changed to be copied into sysspace by the
 * syscall in the future, once issues with various consumers of the
 * attribute code have raised their hands.
 *
 * Currently this is used only by UFS Extended Attributes.
 */
int
extattrctl(td, uap)
	struct thread *td;
	struct extattrctl_args *uap;
{
	struct vnode *filename_vp;
	struct nameidata nd;
	struct mount *mp, *mp_writable;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	/*
	 * SCARG(uap, attrname) not always defined.  We check again later
	 * when we invoke the VFS call so as to pass in NULL there if needed.
	 */
	if (SCARG(uap, attrname) != NULL) {
		error = copyinstr(SCARG(uap, attrname), attrname,
		    EXTATTR_MAXNAMELEN, NULL);
		if (error)
			return (error);
	}

	/*
	 * SCARG(uap, filename) not always defined.  If it is, grab
	 * a vnode lock, which VFS_EXTATTRCTL() will later release.
	 */
	filename_vp = NULL;
	if (SCARG(uap, filename) != NULL) {
		NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
		    SCARG(uap, filename), td);
		if ((error = namei(&nd)) != 0)
			return (error);
		filename_vp = nd.ni_vp;
		NDFREE(&nd, NDF_NO_VP_RELE | NDF_NO_VP_UNLOCK);
	}

	/* SCARG(uap, path) always defined. */
	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0) {
		if (filename_vp != NULL)
			vput(filename_vp);
		return (error);
	}
	mp = nd.ni_vp->v_mount;
	error = vn_start_write(nd.ni_vp, &mp_writable, V_WAIT | PCATCH);
	NDFREE(&nd, 0);
	if (error) {
		if (filename_vp != NULL)
			vput(filename_vp);
		return (error);
	}

	if (SCARG(uap, attrname) != NULL) {
		error = VFS_EXTATTRCTL(mp, SCARG(uap, cmd), filename_vp,
		    SCARG(uap, attrnamespace), attrname, td);
	} else {
		error = VFS_EXTATTRCTL(mp, SCARG(uap, cmd), filename_vp,
		    SCARG(uap, attrnamespace), NULL, td);
	}

	vn_finished_write(mp_writable);
	/*
	 * VFS_EXTATTRCTL will have unlocked, but not de-ref'd,
	 * filename_vp, so vrele it if it is defined.
	 */
	if (filename_vp != NULL)
		vrele(filename_vp);

	return (error);
}

/*-
 * Set a named extended attribute on a file or directory
 * 
 * Arguments: unlocked vnode "vp", attribute namespace "attrnamespace",
 *            kernelspace string pointer "attrname", userspace buffer
 *            pointer "data", buffer length "nbytes", thread "td".
 * Returns: 0 on success, an error number otherwise
 * Locks: none
 * References: vp must be a valid reference for the duration of the call
 */
static int
extattr_set_vp(struct vnode *vp, int attrnamespace, const char *attrname,
    void *data, size_t nbytes, struct thread *td)
{
	struct mount *mp;
	struct uio auio;
	struct iovec aiov;
	ssize_t cnt;
	int error;

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);

	aiov.iov_base = data;
	aiov.iov_len = nbytes;
	auio.uio_iov = &aiov;
	auio.uio_iovcnt = 1;
	auio.uio_offset = 0;
	if (nbytes > INT_MAX) {
		error = EINVAL;
		goto done;
	}
	auio.uio_resid = nbytes;
	auio.uio_rw = UIO_WRITE;
	auio.uio_segflg = UIO_USERSPACE;
	auio.uio_td = td;
	cnt = nbytes;

	error = VOP_SETEXTATTR(vp, attrnamespace, attrname, &auio,
	    td->td_ucred, td);
	cnt -= auio.uio_resid;
	td->td_retval[0] = cnt;

done:
	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return (error);
}

int
extattr_set_file(td, uap)
	struct thread *td;
	struct extattr_set_file_args *uap;
{
	struct nameidata nd;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	    NULL);
	if (error)
		return (error);

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);

	error = extattr_set_vp(nd.ni_vp, SCARG(uap, attrnamespace), attrname,
	    SCARG(uap, data), SCARG(uap, nbytes), td);

	vrele(nd.ni_vp);
	return (error);
}

int
extattr_set_fd(td, uap)
	struct thread *td;
	struct extattr_set_fd_args *uap;
{
	struct file *fp;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	    NULL);
	if (error)
		return (error);

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);

	error = extattr_set_vp((struct vnode *)fp->f_data,
	    SCARG(uap, attrnamespace), attrname, SCARG(uap, data),
	    SCARG(uap, nbytes), td);
	fdrop(fp, td);

	return (error);
}

/*-
 * Get a named extended attribute on a file or directory
 * 
 * Arguments: unlocked vnode "vp", attribute namespace "attrnamespace",
 *            kernelspace string pointer "attrname", userspace buffer
 *            pointer "data", buffer length "nbytes", thread "td".
 * Returns: 0 on success, an error number otherwise
 * Locks: none
 * References: vp must be a valid reference for the duration of the call
 */
static int
extattr_get_vp(struct vnode *vp, int attrnamespace, const char *attrname,
    void *data, size_t nbytes, struct thread *td)
{
	struct uio auio;
	struct iovec aiov;
	ssize_t cnt;
	size_t size;
	int error;

	VOP_LEASE(vp, td, td->td_ucred, LEASE_READ);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);

	/*
	 * Slightly unusual semantics: if the user provides a NULL data
	 * pointer, they don't want to receive the data, just the
	 * maximum read length.
	 */
	if (data != NULL) {
		aiov.iov_base = data;
		aiov.iov_len = nbytes;
		auio.uio_iov = &aiov;
		auio.uio_offset = 0;
		if (nbytes > INT_MAX) {
			error = EINVAL;
			goto done;
		}
		auio.uio_resid = nbytes;
		auio.uio_rw = UIO_READ;
		auio.uio_segflg = UIO_USERSPACE;
		auio.uio_td = td;
		cnt = nbytes;
		error = VOP_GETEXTATTR(vp, attrnamespace, attrname, &auio,
		    NULL, td->td_ucred, td);
		cnt -= auio.uio_resid;
		td->td_retval[0] = cnt;
	} else {
		error = VOP_GETEXTATTR(vp, attrnamespace, attrname, NULL,
		    &size, td->td_ucred, td);
		td->td_retval[0] = size;
	}
done:
	VOP_UNLOCK(vp, 0, td);
	return (error);
}

int
extattr_get_file(td, uap)
	struct thread *td;
	struct extattr_get_file_args *uap;
{
	struct nameidata nd;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	    NULL);
	if (error)
		return (error);

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return (error);
	NDFREE(&nd, NDF_ONLY_PNBUF);

	error = extattr_get_vp(nd.ni_vp, SCARG(uap, attrnamespace), attrname,
	    SCARG(uap, data), SCARG(uap, nbytes), td);

	vrele(nd.ni_vp);
	return (error);
}

int
extattr_get_fd(td, uap)
	struct thread *td;
	struct extattr_get_fd_args *uap;
{
	struct file *fp;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	    NULL);
	if (error)
		return (error);

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);

	error = extattr_get_vp((struct vnode *)fp->f_data,
	    SCARG(uap, attrnamespace), attrname, SCARG(uap, data),
	    SCARG(uap, nbytes), td);

	fdrop(fp, td);
	return (error);
}

/*
 * extattr_delete_vp(): Delete a named extended attribute on a file or
 *                      directory
 * 
 * Arguments: unlocked vnode "vp", attribute namespace "attrnamespace",
 *            kernelspace string pointer "attrname", proc "p"
 * Returns: 0 on success, an error number otherwise
 * Locks: none
 * References: vp must be a valid reference for the duration of the call
 */
static int
extattr_delete_vp(struct vnode *vp, int attrnamespace, const char *attrname,
    struct thread *td)
{
	struct mount *mp;
	int error;

	if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0)
		return (error);
	VOP_LEASE(vp, td, td->td_ucred, LEASE_WRITE);
	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);

	error = VOP_SETEXTATTR(vp, attrnamespace, attrname, NULL, td->td_ucred,
	    td);

	VOP_UNLOCK(vp, 0, td);
	vn_finished_write(mp);
	return (error);
}

int
extattr_delete_file(td, uap)
	struct thread *td;
	struct extattr_delete_file_args *uap;
{
	struct nameidata nd;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	     NULL);
	if (error)
		return(error);

	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), td);
	if ((error = namei(&nd)) != 0)
		return(error);
	NDFREE(&nd, NDF_ONLY_PNBUF);

	error = extattr_delete_vp(nd.ni_vp, SCARG(uap, attrnamespace),
	    attrname, td);

	vrele(nd.ni_vp);
	return(error);
}

int
extattr_delete_fd(td, uap)
	struct thread *td;
	struct extattr_delete_fd_args *uap;
{
	struct file *fp;
	struct vnode *vp;
	char attrname[EXTATTR_MAXNAMELEN];
	int error;

	error = copyinstr(SCARG(uap, attrname), attrname, EXTATTR_MAXNAMELEN,
	    NULL);
	if (error)
		return (error);

	if ((error = getvnode(td->td_proc->p_fd, SCARG(uap, fd), &fp)) != 0)
		return (error);
	vp = (struct vnode *)fp->f_data;

	error = extattr_delete_vp((struct vnode *)fp->f_data,
	    SCARG(uap, attrnamespace), attrname, td);

	fdrop(fp, td);
	return (error);
}