shithub: riscv

ref: f88a55e79b5bf656e7f9578d1318a955b9a4963a
dir: /sys/src/cmd/diff/test/diff-t9.expected/

View raw version
--- diff-t9.l
+++ diff-t9.r
@@ -1,4 +1,5 @@
-/*	$NetBSD: vfs_syscalls.c,v 1.57 1995/10/07 06:28:51 mycroft Exp $	*/
+/*	$OpenBSD: t9.2,v 1.2 2013/12/01 16:40:56 krw Exp $	*/
+/*	$NetBSD: vfs_syscalls.c,v 1.71 1996/04/23 10:29:02 mycroft Exp $	*/
 
 /*
  * Copyright (c) 1989, 1993
@@ -17,11 +18,7 @@
  * 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
+ * 3. 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.
  *
@@ -53,13 +50,30 @@
 #include <sys/uio.h>
 #include <sys/malloc.h>
 #include <sys/dirent.h>
+#include <sys/extattr.h>
 
 #include <sys/syscallargs.h>
 
-#include <vm/vm.h>
+#include <uvm/uvm_extern.h>
 #include <sys/sysctl.h>
 
-static int change_dir __P((struct nameidata *ndp, struct proc *p));
+extern int suid_clear;
+int	usermount = 0;		/* sysctl: by default, users may not mount */
+
+static int change_dir(struct nameidata *, struct proc *);
+
+void checkdirs(struct vnode *);
+
+/*
+ * Redirection info so we don't have to include the union fs routines in
+ * the kernel directly.  This way, we can build unionfs as an LKM.  The
+ * pointer gets filled in later, when we modload the LKM, or when the
+ * compiled-in unionfs code gets initialized.  For now, we just set
+ * it to a stub routine.
+ */
+
+int (*union_check_p)(struct proc *, struct vnode **,
+    struct file *, struct uio, int *) = NULL;
 
 /*
  * Virtual File System System Calls
@@ -69,6 +83,7 @@
  * Mount a file system.
  */
 /* ARGSUSED */
+int
 sys_mount(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -78,22 +93,36 @@
 		syscallarg(char *) type;
 		syscallarg(char *) path;
 		syscallarg(int) flags;
-		syscallarg(caddr_t) data;
+		syscallarg(void *) data;
 	} */ *uap = v;
 	register struct vnode *vp;
 	register struct mount *mp;
-	int error, flag;
-	u_long fsindex;
+	int error, flag = 0;
+#ifdef COMPAT_43
+	u_long fstypenum = 0;
+#endif
 	char fstypename[MFSNAMELEN];
+	char fspath[MNAMELEN];
 	struct vattr va;
 	struct nameidata nd;
+	struct vfsconf *vfsp;
+	struct timeval tv;
+
+	if (usermount == 0 && (error = suser(p->p_ucred, &p->p_acflag)))
+		return (error);
+
+	/*
+	 * Mount points must fit in MNAMELEN, not MAXPATHLEN.
+	 */
+	error = copyinstr(SCARG(uap, path), fspath, MNAMELEN, NULL);
+	if (error)
+		return(error);
 
 	/*
 	 * Get vnode to be covered
 	 */
-	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
-	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, fspath, p);
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	if (SCARG(uap, flags) & MNT_UPDATE) {
@@ -113,7 +142,7 @@
 			return (EOPNOTSUPP);	/* Needs translation */
 		}
 		mp->mnt_flag |=
-		    SCARG(uap, flags) & (MNT_RELOAD | MNT_FORCE | MNT_UPDATE);
+		    SCARG(uap, flags) & (MNT_RELOAD | MNT_UPDATE);
 		/*
 		 * Only root, or the user that did the original mount is
 		 * permitted to update it.
@@ -134,7 +163,11 @@
 			}
 			SCARG(uap, flags) |= MNT_NOSUID | MNT_NODEV;
 		}
-		VOP_UNLOCK(vp);
+		if ((error = vfs_busy(mp, LK_NOWAIT, 0, p)) != 0) {
+			vput(vp);
+			return (error);
+		}
+		VOP_UNLOCK(vp, 0, p);
 		goto update;
 	}
 	/*
@@ -143,7 +176,7 @@
 	 */
 	if ((error = VOP_GETATTR(vp, &va, p->p_ucred, p)) ||
 	    (va.va_uid != p->p_ucred->cr_uid &&
-	     (error = suser(p->p_ucred, &p->p_acflag)))) {
+	    (error = suser(p->p_ucred, &p->p_acflag)))) {
 		vput(vp);
 		return (error);
 	}
@@ -158,40 +191,47 @@
 		}
 		SCARG(uap, flags) |= MNT_NOSUID | MNT_NODEV;
 	}
-	if (error = vinvalbuf(vp, V_SAVE, p->p_ucred, p, 0, 0))
+	if ((error = vinvalbuf(vp, V_SAVE, p->p_ucred, p, 0, 0)) != 0)
 		return (error);
 	if (vp->v_type != VDIR) {
 		vput(vp);
 		return (ENOTDIR);
 	}
-	if (error = copyinstr(SCARG(uap, type), fstypename, MFSNAMELEN,
-	    (size_t *)0)) {
-#if defined(COMPAT_09) || defined(COMPAT_43)
+	error = copyinstr(SCARG(uap, type), fstypename, MFSNAMELEN, NULL);
+	if (error) {
+#ifdef COMPAT_43
 		/*
 		 * Historically filesystem types were identified by number.
 		 * If we get an integer for the filesystem type instead of a
 		 * string, we check to see if it matches one of the historic
 		 * filesystem types.
-		 */     
-		fsindex = (u_long)SCARG(uap, type);
-		if (fsindex >= nvfssw || vfssw[fsindex] == NULL) {
+		 */
+		fstypenum = (u_long)SCARG(uap, type);
+
+		for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next)
+			if (vfsp->vfc_typenum == fstypenum)
+				break;
+		if (vfsp == NULL) {
 			vput(vp);
 			return (ENODEV);
 		}
-		strncpy(fstypename, vfssw[fsindex]->vfs_name, MFSNAMELEN);
+		strncpy(fstypename, vfsp->vfc_name, MFSNAMELEN);
+
 #else
 		vput(vp);
 		return (error);
 #endif
 	}
-	for (fsindex = 0; fsindex < nvfssw; fsindex++)
-		if (vfssw[fsindex] != NULL &&
-		    !strncmp(vfssw[fsindex]->vfs_name, fstypename, MFSNAMELEN))
+	for (vfsp = vfsconf; vfsp; vfsp = vfsp->vfc_next) {
+		if (!strcmp(vfsp->vfc_name, fstypename))
 			break;
-	if (fsindex >= nvfssw) {
+	}
+
+	if (vfsp == NULL) {
 		vput(vp);
-		return (ENODEV);
+		return (EOPNOTSUPP);
 	}
+
 	if (vp->v_mountedhere != NULL) {
 		vput(vp);
 		return (EBUSY);
@@ -203,15 +243,14 @@
 	mp = (struct mount *)malloc((u_long)sizeof(struct mount),
 		M_MOUNT, M_WAITOK);
 	bzero((char *)mp, (u_long)sizeof(struct mount));
-	mp->mnt_op = vfssw[fsindex];
-	if (error = vfs_lock(mp)) {
-		free((caddr_t)mp, M_MOUNT);
-		vput(vp);
-		return (error);
-	}
-	/* Do this early in case we block later. */
-	vfssw[fsindex]->vfs_refcount++;
-	vp->v_mountedhere = mp;
+	lockinit(&mp->mnt_lock, PVFS, "vfslock", 0, 0);
+	/* This error never happens, but it makes auditing easier */
+	if ((error = vfs_busy(mp, LK_NOWAIT, 0, p)))
+		return (error);
+	mp->mnt_op = vfsp->vfc_vfsops;
+	mp->mnt_vfc = vfsp;
+	mp->mnt_flag |= (vfsp->vfc_flags & MNT_VISFLAGMASK);
+	strncpy(mp->mnt_stat.f_fstypename, vfsp->vfc_name, MFSNAMELEN);
 	mp->mnt_vnodecovered = vp;
 	mp->mnt_stat.f_owner = p->p_ucred->cr_uid;
 update:
@@ -223,13 +262,19 @@
 	else if (mp->mnt_flag & MNT_RDONLY)
 		mp->mnt_flag |= MNT_WANTRDWR;
 	mp->mnt_flag &=~ (MNT_NOSUID | MNT_NOEXEC | MNT_NODEV |
-	    MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC);
+	    MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC | MNT_SOFTDEP |
+	    MNT_NOATIME | MNT_FORCE);
 	mp->mnt_flag |= SCARG(uap, flags) & (MNT_NOSUID | MNT_NOEXEC |
-	    MNT_NODEV | MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC);
+	    MNT_NODEV | MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC |
+	    MNT_SOFTDEP | MNT_NOATIME | MNT_FORCE);
 	/*
 	 * Mount the filesystem.
 	 */
 	error = VFS_MOUNT(mp, SCARG(uap, path), SCARG(uap, data), &nd, p);
+	if (!error) {
+		microtime(&tv);
+		mp->mnt_stat.f_ctime = tv.tv_sec;
+	}
 	if (mp->mnt_flag & MNT_UPDATE) {
 		vrele(vp);
 		if (mp->mnt_flag & MNT_WANTRDWR)
@@ -238,23 +283,42 @@
 		    (MNT_UPDATE | MNT_RELOAD | MNT_FORCE | MNT_WANTRDWR);
 		if (error)
 			mp->mnt_flag = flag;
+
+ 		if ((mp->mnt_flag & MNT_RDONLY) == 0) {
+ 			if (mp->mnt_syncer == NULL)
+ 				error = vfs_allocate_syncvnode(mp);
+ 		} else {
+ 			if (mp->mnt_syncer != NULL)
+ 				vgone(mp->mnt_syncer);
+ 			mp->mnt_syncer = NULL;
+ 		}
+
+		vfs_unbusy(mp, p);
 		return (error);
 	}
+
+	vp->v_mountedhere = mp;
+
 	/*
 	 * Put the new filesystem on the mount list after root.
 	 */
 	cache_purge(vp);
 	if (!error) {
+		vfsp->vfc_refcount++;
+		simple_lock(&mountlist_slock);
 		TAILQ_INSERT_TAIL(&mountlist, mp, mnt_list);
+		simple_unlock(&mountlist_slock);
 		checkdirs(vp);
-		VOP_UNLOCK(vp);
-		vfs_unlock(mp);
+		VOP_UNLOCK(vp, 0, p);
+ 		if ((mp->mnt_flag & MNT_RDONLY) == 0)
+ 			error = vfs_allocate_syncvnode(mp);
+		vfs_unbusy(mp, p);
 		(void) VFS_STATFS(mp, &mp->mnt_stat, p);
-		error = VFS_START(mp, 0, p);
+		if ((error = VFS_START(mp, 0, p)) != 0)
+			vrele(vp);
 	} else {
 		mp->mnt_vnodecovered->v_mountedhere = (struct mount *)0;
-		vfssw[fsindex]->vfs_refcount--;
-		vfs_unlock(mp);
+		vfs_unbusy(mp, p);
 		free((caddr_t)mp, M_MOUNT);
 		vput(vp);
 	}
@@ -266,6 +330,7 @@
  * or root directory onto which the new filesystem has just been
  * mounted. If so, replace them with the new mount point.
  */
+void
 checkdirs(olddp)
 	struct vnode *olddp;
 {
@@ -277,7 +342,7 @@
 		return;
 	if (VFS_ROOT(olddp->v_mountedhere, &newdp))
 		panic("mount: lost mount");
-	for (p = allproc.lh_first; p != 0; p = p->p_list.le_next) {
+	for (p = LIST_FIRST(&allproc); p != 0; p = LIST_NEXT(p, p_list)) {
 		fdp = p->p_fd;
 		if (fdp->fd_cdir == olddp) {
 			vrele(fdp->fd_cdir);
@@ -305,6 +370,7 @@
  * not special file (as before).
  */
 /* ARGSUSED */
+int
 sys_unmount(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -321,7 +387,7 @@
 
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	mp = vp->v_mount;
@@ -352,50 +418,77 @@
 		return (EINVAL);
 	}
 	vput(vp);
-	return (dounmount(mp, SCARG(uap, flags), p));
+
+	if (vfs_busy(mp, LK_EXCLUSIVE, NULL, p))
+		return (EBUSY);
+
+	return (dounmount(mp, SCARG(uap, flags), p, vp));
 }
 
 /*
  * Do the actual file system unmount.
  */
-dounmount(mp, flags, p)
-	register struct mount *mp;
-	int flags;
-	struct proc *p;
+int
+dounmount(struct mount *mp, int flags, struct proc *p, struct vnode *olddp)
 {
 	struct vnode *coveredvp;
+	struct proc *p2;
 	int error;
+	int hadsyncer = 0;
 
-	coveredvp = mp->mnt_vnodecovered;
-	if (vfs_busy(mp))
-		return (EBUSY);
-	mp->mnt_flag |= MNT_UNMOUNT;
-	if (error = vfs_lock(mp))
-		return (error);
-
-	mp->mnt_flag &=~ MNT_ASYNC;
-	vnode_pager_umount(mp);	/* release cached vnodes */
-	cache_purgevfs(mp);	/* remove cache entries for this file sys */
-	if ((error = VFS_SYNC(mp, MNT_WAIT, p->p_ucred, p)) == 0 ||
-	    (flags & MNT_FORCE))
-		error = VFS_UNMOUNT(mp, flags, p);
-	mp->mnt_flag &= ~MNT_UNMOUNT;
-	vfs_unbusy(mp);
-	if (error) {
-		vfs_unlock(mp);
-	} else {
-		TAILQ_REMOVE(&mountlist, mp, mnt_list);
-		if (coveredvp != NULLVP) {
-			vrele(coveredvp);
-			coveredvp->v_mountedhere = (struct mount *)0;
-		}
-		mp->mnt_op->vfs_refcount--;
-		vfs_unlock(mp);
-		if (mp->mnt_vnodelist.lh_first != NULL)
-			panic("unmount: dangling vnode");
-		free((caddr_t)mp, M_MOUNT);
-	}
-	return (error);
+ 	mp->mnt_flag &=~ MNT_ASYNC;
+ 	cache_purgevfs(mp);	/* remove cache entries for this file sys */
+ 	if (mp->mnt_syncer != NULL) {
+		hadsyncer = 1;
+ 		vgone(mp->mnt_syncer);
+		mp->mnt_syncer = NULL;
+	}
+	if (((mp->mnt_flag & MNT_RDONLY) ||
+	    (error = VFS_SYNC(mp, MNT_WAIT, p->p_ucred, p)) == 0) ||
+ 	    (flags & MNT_FORCE))
+ 		error = VFS_UNMOUNT(mp, flags, p);
+	simple_lock(&mountlist_slock);
+ 	if (error) {
+ 		if ((mp->mnt_flag & MNT_RDONLY) == 0 && hadsyncer)
+ 			(void) vfs_allocate_syncvnode(mp);
+		lockmgr(&mp->mnt_lock, LK_RELEASE | LK_INTERLOCK,
+		    &mountlist_slock, p);
+		return (error);
+	}
+	TAILQ_REMOVE(&mountlist, mp, mnt_list);
+	if ((coveredvp = mp->mnt_vnodecovered) != NULLVP) {
+		if (olddp) {
+			/* 
+			 * Try to put processes back in a real directory
+			 * after a forced unmount.
+			 * XXX We're not holding a ref on olddp, which may
+			 * change, so compare id numbers.
+			 */
+			LIST_FOREACH(p2, &allproc, p_list) {
+				struct filedesc *fdp = p2->p_fd;
+				if (fdp->fd_cdir &&
+				    fdp->fd_cdir->v_id == olddp->v_id) {
+					vrele(fdp->fd_cdir);
+					vref(coveredvp);
+					fdp->fd_cdir = coveredvp;
+				}
+				if (fdp->fd_rdir &&
+				    fdp->fd_rdir->v_id == olddp->v_id) {
+					vrele(fdp->fd_rdir);
+					vref(coveredvp);
+					fdp->fd_rdir = coveredvp;
+				}
+			}
+		}
+		coveredvp->v_mountedhere = NULL;
+ 		vrele(coveredvp);
+ 	}
+	mp->mnt_vfc->vfc_refcount--;
+	if (mp->mnt_vnodelist.lh_first != NULL)
+		panic("unmount: dangling vnode");
+	lockmgr(&mp->mnt_lock, LK_RELEASE | LK_INTERLOCK, &mountlist_slock, p);
+	free((caddr_t)mp, M_MOUNT);
+	return (0);
 }
 
 /*
@@ -407,6 +500,7 @@
 #endif
 
 /* ARGSUSED */
+int
 sys_sync(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -415,31 +509,23 @@
 	register struct mount *mp, *nmp;
 	int asyncflag;
 
-	for (mp = mountlist.cqh_first; mp != (void *)&mountlist; mp = nmp) {
-		/*
-		 * Get the next pointer in case we hang on vfs_busy
-		 * while we are being unmounted.
-		 */
-		nmp = mp->mnt_list.cqe_next;
-		/*
-		 * The lock check below is to avoid races with mount
-		 * and unmount.
-		 */
-		if ((mp->mnt_flag & (MNT_MLOCK|MNT_RDONLY|MNT_MPBUSY)) == 0 &&
-		    !vfs_busy(mp)) {
+	simple_lock(&mountlist_slock);
+	TAILQ_FOREACH_REVERSE_SAFE(mp, &mountlist, mnt_list, nmp) {
+		if (vfs_busy(mp, LK_NOWAIT, &mountlist_slock, p))
+			continue;
+		if ((mp->mnt_flag & MNT_RDONLY) == 0) {
 			asyncflag = mp->mnt_flag & MNT_ASYNC;
 			mp->mnt_flag &= ~MNT_ASYNC;
+			uvm_vnp_sync(mp);
 			VFS_SYNC(mp, MNT_NOWAIT, p->p_ucred, p);
 			if (asyncflag)
 				mp->mnt_flag |= MNT_ASYNC;
-			/*
-			 * Get the next pointer again, as the next filesystem
-			 * might have been unmounted while we were sync'ing.
-			 */
-			nmp = mp->mnt_list.cqe_next;
-			vfs_unbusy(mp);
 		}
+		simple_lock(&mountlist_slock);
+		vfs_unbusy(mp, p);
 	}
+	simple_unlock(&mountlist_slock);
+
 #ifdef DEBUG
 	if (syncprt)
 		vfs_bufstats();
@@ -451,6 +537,7 @@
  * Change filesystem quotas.
  */
 /* ARGSUSED */
+int
 sys_quotactl(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -467,7 +554,7 @@
 	struct nameidata nd;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	mp = nd.ni_vp->v_mount;
 	vrele(nd.ni_vp);
@@ -479,6 +566,7 @@
  * Get filesystem statistics.
  */
 /* ARGSUSED */
+int
 sys_statfs(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -492,16 +580,27 @@
 	register struct statfs *sp;
 	int error;
 	struct nameidata nd;
+	struct statfs sb;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	mp = nd.ni_vp->v_mount;
 	sp = &mp->mnt_stat;
 	vrele(nd.ni_vp);
-	if (error = VFS_STATFS(mp, sp, p))
+	if ((error = VFS_STATFS(mp, sp, p)) != 0)
 		return (error);
 	sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
+#if notyet
+	if (mp->mnt_flag & MNT_SOFTDEP)
+		sp->f_eflags = STATFS_SOFTUPD;
+#endif
+	/* Don't let non-root see filesystem id (for NFS security) */
+	if (suser(p->p_ucred, &p->p_acflag)) {
+		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)));
 }
 
@@ -509,33 +608,48 @@
  * Get filesystem statistics.
  */
 /* ARGSUSED */
+int
 sys_fstatfs(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_fstatfs_args /* {
+	struct sys_fstatfs_args /* {
 		syscallarg(int) fd;
 		syscallarg(struct statfs *) buf;
 	} */ *uap = v;
 	struct file *fp;
 	struct mount *mp;
-	register struct statfs *sp;
+	struct statfs *sp;
 	int error;
+	struct statfs sb;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	mp = ((struct vnode *)fp->f_data)->v_mount;
 	sp = &mp->mnt_stat;
-	if (error = VFS_STATFS(mp, sp, p))
+	error = VFS_STATFS(mp, sp, p);
+	FRELE(fp);
+	if (error)
 		return (error);
 	sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
+#if notyet
+	if (mp->mnt_flag & MNT_SOFTDEP)
+		sp->f_eflags = STATFS_SOFTUPD;
+#endif
+	/* Don't let non-root see filesystem id (for NFS security) */
+	if (suser(p->p_ucred, &p->p_acflag)) {
+		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.
  */
+int
 sys_getfsstat(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -543,37 +657,59 @@
 {
 	register struct sys_getfsstat_args /* {
 		syscallarg(struct statfs *) buf;
-		syscallarg(long) bufsize;
+		syscallarg(size_t) bufsize;
 		syscallarg(int) flags;
 	} */ *uap = v;
-	register struct mount *mp, *nmp;
+	register struct mount *mp *nmp;
 	register struct statfs *sp;
+	struct statfs sb;
 	caddr_t sfsp;
-	long count, maxcount, error;
+	size_t count, maxcount;
+	int error, flags = SCARG(uap, flags);
 
 	maxcount = SCARG(uap, bufsize) / sizeof(struct statfs);
 	sfsp = (caddr_t)SCARG(uap, buf);
-	for (count = 0,
-	     mp = mountlist.cqh_first; mp != (void *)&mountlist; mp = nmp) {
-		nmp = mp->mnt_list.cqe_next;
-		if (sfsp && count < maxcount &&
-		    ((mp->mnt_flag & MNT_MLOCK) == 0)) {
+	count = 0;
+	simple_lock(&mountlist_slock);
+	TAILQ_FOREACH_REVERSE_SAFE(mp, &mountlist, mnt_list, nmp) {
+		if (vfs_busy(mp, LK_NOWAIT, &mountlist_slock, p))
+			continue;
+		if (sfsp && count < maxcount) {
 			sp = &mp->mnt_stat;
-			/*
-			 * If MNT_NOWAIT is specified, do not refresh the
-			 * fsstat cache. MNT_WAIT overrides MNT_NOWAIT.
-			 */
-			if (((SCARG(uap, flags) & MNT_NOWAIT) == 0 ||
-			    (SCARG(uap, flags) & MNT_WAIT)) &&
-			    (error = VFS_STATFS(mp, sp, p)))
-				continue;
+
+			/* Refresh stats unless MNT_NOWAIT is specified */
+			if (flags != MNT_NOWAIT &&
+			    flags != MNT_LAZY &&
+			    (flags == MNT_WAIT ||
+			    flags == 0) &&
+			    (error = VFS_STATFS(mp, sp, p))) {
+				simple_lock(&mountlist_slock);
+				vfs_unbusy(mp, p);
+ 				continue;
+			}
+
 			sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
-			if (error = copyout((caddr_t)sp, sfsp, sizeof(*sp)))
+#if notyet
+			if (mp->mnt_flag & MNT_SOFTDEP)
+				sp->f_eflags = STATFS_SOFTUPD;
+#endif
+			if (suser(p->p_ucred, &p->p_acflag)) {
+				bcopy((caddr_t)sp, (caddr_t)&sb, sizeof(sb));
+				sb.f_fsid.val[0] = sb.f_fsid.val[1] = 0;
+				sp = &sb;
+			}
+			error = copyout((caddr_t)sp, sfsp, sizeof(*sp));
+			if (error) {
+				vfs_unbusy(mp, p);
 				return (error);
+			}
 			sfsp += sizeof(*sp);
 		}
 		count++;
+		simple_lock(&mountlist_slock);
+		vfs_unbusy(mp, p);
 	}
+	simple_unlock(&mountlist_slock);
 	if (sfsp && count > maxcount)
 		*retval = maxcount;
 	else
@@ -585,6 +721,7 @@
  * Change current working directory to a given file descriptor.
  */
 /* ARGSUSED */
+int
 sys_fchdir(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -593,37 +730,38 @@
 	struct sys_fchdir_args /* {
 		syscallarg(int) fd;
 	} */ *uap = v;
-	register struct filedesc *fdp = p->p_fd;
+	struct filedesc *fdp = p->p_fd;
 	struct vnode *vp, *tdp;
 	struct mount *mp;
 	struct file *fp;
 	int error;
 
-	if (error = getvnode(fdp, SCARG(uap, fd), &fp))
+	if ((error = getvnode(fdp, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	vp = (struct vnode *)fp->f_data;
 	VREF(vp);
-	VOP_LOCK(vp);
+	FRELE(fp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_type != VDIR)
 		error = ENOTDIR;
 	else
 		error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p);
+
 	while (!error && (mp = vp->v_mountedhere) != NULL) {
-		if (mp->mnt_flag & MNT_MLOCK) {
-			mp->mnt_flag |= MNT_MWAIT;
-			sleep((caddr_t)mp, PVFS);
+		if (vfs_busy(mp, 0, 0, p))
 			continue;
-		}
-		if (error = VFS_ROOT(mp, &tdp))
+		error = VFS_ROOT(mp, &tdp);
+		vfs_unbusy(mp, p);
+		if (error)
 			break;
 		vput(vp);
 		vp = tdp;
 	}
-	VOP_UNLOCK(vp);
 	if (error) {
-		vrele(vp);
+		vput(vp);
 		return (error);
 	}
+	VOP_UNLOCK(vp, 0, p);
 	vrele(fdp->fd_cdir);
 	fdp->fd_cdir = vp;
 	return (0);
@@ -633,6 +771,7 @@
  * Change current working directory (``.'').
  */
 /* ARGSUSED */
+int
 sys_chdir(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -647,7 +786,7 @@
 
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = change_dir(&nd, p))
+	if ((error = change_dir(&nd, p)) != 0)
 		return (error);
 	vrele(fdp->fd_cdir);
 	fdp->fd_cdir = nd.ni_vp;
@@ -658,6 +797,7 @@
  * Change notion of root (``/'') directory.
  */
 /* ARGSUSED */
+int
 sys_chroot(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -670,14 +810,22 @@
 	int error;
 	struct nameidata nd;
 
-	if (error = suser(p->p_ucred, &p->p_acflag))
+	if ((error = suser(p->p_ucred, &p->p_acflag)) != 0)
 		return (error);
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = change_dir(&nd, p))
+	if ((error = change_dir(&nd, p)) != 0)
 		return (error);
-	if (fdp->fd_rdir != NULL)
+	if (fdp->fd_rdir != NULL) {
+		/*
+		 * A chroot() done inside a changed root environment does
+		 * an automatic chdir to avoid the out-of-tree experience.
+		 */
 		vrele(fdp->fd_rdir);
+		vrele(fdp->fd_cdir);
+		VREF(nd.ni_vp);
+		fdp->fd_cdir = nd.ni_vp;
+	}
 	fdp->fd_rdir = nd.ni_vp;
 	return (0);
 }
@@ -693,16 +841,17 @@
 	struct vnode *vp;
 	int error;
 
-	if (error = namei(ndp))
+	if ((error = namei(ndp)) != 0)
 		return (error);
 	vp = ndp->ni_vp;
 	if (vp->v_type != VDIR)
 		error = ENOTDIR;
 	else
 		error = VOP_ACCESS(vp, VEXEC, p->p_ucred, p);
-	VOP_UNLOCK(vp);
 	if (error)
-		vrele(vp);
+		vput(vp);
+	else
+		VOP_UNLOCK(vp, 0, p);
 	return (error);
 }
 
@@ -710,45 +859,50 @@
  * Check permissions, allocate an open file structure,
  * and call the device open routine if any.
  */
+int
 sys_open(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_open_args /* {
+	struct sys_open_args /* {
 		syscallarg(char *) path;
 		syscallarg(int) flags;
 		syscallarg(int) mode;
 	} */ *uap = v;
-	register struct filedesc *fdp = p->p_fd;
-	register struct file *fp;
-	register struct vnode *vp;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp;
+	struct vattr vattr;
 	int flags, cmode;
-	struct file *nfp;
-	int type, indx, error;
+	int type, indx, error, localtrunc = 0;
 	struct flock lf;
 	struct nameidata nd;
-	extern struct fileops vnops;
 
-	if (error = falloc(p, &nfp, &indx))
+	if ((error = falloc(p, &fp, &indx)) != 0)
 		return (error);
-	fp = nfp;
+
 	flags = FFLAGS(SCARG(uap, flags));
 	cmode = ((SCARG(uap, mode) &~ fdp->fd_cmask) & ALLPERMS) &~ S_ISTXT;
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
 	p->p_dupfd = -indx - 1;			/* XXX check for fdopen */
-	if (error = vn_open(&nd, flags, cmode)) {
-		ffree(fp);
+	if ((flags & O_TRUNC) && (flags & (O_EXLOCK | O_SHLOCK))) {
+		localtrunc = 1;
+		flags &= ~O_TRUNC;	/* Must do truncate ourselves */
+	}
+	if ((error = vn_open(&nd, flags, cmode)) != 0) {
 		if ((error == ENODEV || error == ENXIO) &&
 		    p->p_dupfd >= 0 &&			/* XXX from fdopen */
 		    (error =
 			dupfdopen(fdp, indx, p->p_dupfd, flags, error)) == 0) {
+			closef(fp, p);
 			*retval = indx;
 			return (0);
 		}
 		if (error == ERESTART)
 			error = EINTR;
-		fdp->fd_ofiles[indx] = NULL;
+		fdremove(fdp, indx);
+		closef(fp, p);
 		return (error);
 	}
 	p->p_dupfd = 0;
@@ -768,189 +922,463 @@
 		type = F_FLOCK;
 		if ((flags & FNONBLOCK) == 0)
 			type |= F_WAIT;
-		VOP_UNLOCK(vp);
-		if (error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type)) {
-			(void) vn_close(vp, fp->f_flag, fp->f_cred, p);
-			ffree(fp);
-			fdp->fd_ofiles[indx] = NULL;
+		VOP_UNLOCK(vp, 0, p);
+		error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type);
+		if (error) {
+			/* closef will vn_close the file for us. */
+			fdremove(fdp, indx);
+			closef(fp, p);
 			return (error);
 		}
-		VOP_LOCK(vp);
+		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 		fp->f_flag |= FHASLOCK;
 	}
-	VOP_UNLOCK(vp);
+	if (localtrunc) {
+		VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+		if ((fp->f_flag & FWRITE) == 0)
+			error = EACCES;
+		else if (vp->v_mount->mnt_flag & MNT_RDONLY)
+			error = EROFS;
+		else if (vp->v_type == VDIR)
+			error = EISDIR;
+		else if ((error = vn_writechk(vp)) == 0) {
+			VATTR_NULL(&vattr);
+			vattr.va_size = 0;
+			error = VOP_SETATTR(vp, &vattr, fp->f_cred, p);
+		}
+		if (error) {
+			VOP_UNLOCK(vp, 0, p);
+			/* closef will close the file for us. */
+			fdremove(fdp, indx);
+			closef(fp, p);
+			return (error);
+		}
+	}
+	VOP_UNLOCK(vp, 0, p);
 	*retval = indx;
+	FILE_SET_MATURE(fp);
 	return (0);
 }
 
 /*
- * Create a special file.
+ * Get file handle system call
  */
-/* ARGSUSED */
-sys_mknod(p, v, retval)
+int
+sys_getfh(p, v, retval)
 	struct proc *p;
-	void *v;
+	register void *v;
 	register_t *retval;
 {
-	register struct sys_mknod_args /* {
-		syscallarg(char *) path;
-		syscallarg(int) mode;
-		syscallarg(int) dev;
+	register struct sys_getfh_args /* {
+		syscallarg(char *) fname;
+		syscallarg(fhandle_t *) fhp;
 	} */ *uap = v;
 	register struct vnode *vp;
-	struct vattr vattr;
+	fhandle_t fh;
 	int error;
-	int whiteout;
 	struct nameidata nd;
 
-	if (error = suser(p->p_ucred, &p->p_acflag))
+	/*
+	 * Must be super user
+	 */
+	error = suser(p->p_ucred, &p->p_acflag);
+	if (error)
 		return (error);
-	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
+	    SCARG(uap, fname), p);
+	error = namei(&nd);
+	if (error)
 		return (error);
 	vp = nd.ni_vp;
-	if (vp != NULL)
-		error = EEXIST;
-	else {
-		VATTR_NULL(&vattr);
-		vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ p->p_fd->fd_cmask;
-		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 (!error) {
-		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-		if (whiteout) {
-			error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, CREATE);
-			if (error)
-				VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
-			vput(nd.ni_dvp);
-		} else {
-			error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp,
-						&nd.ni_cnd, &vattr);
-		}
-	} else {
-		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
-		if (nd.ni_dvp == vp)
-			vrele(nd.ni_dvp);
-		else
-			vput(nd.ni_dvp);
-		if (vp)
-			vrele(vp);
-	}
+	bzero((caddr_t)&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((caddr_t)&fh, (caddr_t)SCARG(uap, fhp), sizeof (fh));
 	return (error);
 }
 
 /*
- * Create a named pipe.
+ * Open a file given a file handle.
+ *
+ * Check permissions, allocate an open file structure,
+ * and call the device open routine if any.
  */
-/* ARGSUSED */
-sys_mkfifo(p, v, retval)
+int
+sys_fhopen(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_mkfifo_args /* {
-		syscallarg(char *) path;
-		syscallarg(int) mode;
+	register struct sys_fhopen_args /* {
+		syscallarg(const fhandle_t *) fhp;
+		syscallarg(int) flags;
 	} */ *uap = v;
-	struct vattr vattr;
-	int error;
-	struct nameidata nd;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp = NULL;
+	struct mount *mp;
+	struct ucred *cred = p->p_ucred;
+	int flags;
+	int type, indx, error=0;
+	struct flock lf;
+	struct vattr va;
+	fhandle_t fh;
 
-#ifndef FIFO
-	return (EOPNOTSUPP);
-#else
-	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	/*
+	 * Must be super user
+	 */
+	if ((error = suser(p->p_ucred, &p->p_acflag)))
 		return (error);
-	if (nd.ni_vp != NULL) {
-		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
-		if (nd.ni_dvp == nd.ni_vp)
-			vrele(nd.ni_dvp);
-		else
-			vput(nd.ni_dvp);
-		vrele(nd.ni_vp);
-		return (EEXIST);
-	}
-	VATTR_NULL(&vattr);
-	vattr.va_type = VFIFO;
-	vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ p->p_fd->fd_cmask;
-	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-	return (VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr));
-#endif /* FIFO */
-}
 
-/*
- * Make a hard file link.
- */
-/* ARGSUSED */
-sys_link(p, v, retval)
-	struct proc *p;
-	void *v;
-	register_t *retval;
-{
-	register struct sys_link_args /* {
-		syscallarg(char *) path;
-		syscallarg(char *) link;
-	} */ *uap = v;
-	register struct vnode *vp;
-	struct nameidata nd;
-	int error;
+	flags = FFLAGS(SCARG(uap, flags));
+	if ((flags & (FREAD | FWRITE)) == 0)
+		return (EINVAL);
+	if ((flags & O_CREAT))
+		return (EINVAL);
 
-	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = falloc(p, &fp, &indx)) != 0)
 		return (error);
-	vp = nd.ni_vp;
-	if (vp->v_type != VDIR ||
-	    (error = suser(p->p_ucred, &p->p_acflag)) == 0) {
-		nd.ni_cnd.cn_nameiop = CREATE;
-		nd.ni_cnd.cn_flags = LOCKPARENT;
-		nd.ni_dirp = SCARG(uap, link);
-		if ((error = namei(&nd)) == 0) {
-			if (nd.ni_vp != NULL)
-				error = EEXIST;
-			if (!error) {
-				VOP_LEASE(nd.ni_dvp, p, p->p_ucred,
-				    LEASE_WRITE);
-				VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-				error = VOP_LINK(nd.ni_dvp, vp, &nd.ni_cnd);
-			} else {
-				VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
-				if (nd.ni_dvp == nd.ni_vp)
-					vrele(nd.ni_dvp);
-				else
-					vput(nd.ni_dvp);
-				if (nd.ni_vp)
-					vrele(nd.ni_vp);
-			}
-		}
+
+	if ((error = copyin(SCARG(uap, fhp), &fh, sizeof(fhandle_t))) != 0)
+		goto bad;
+
+	if ((mp = vfs_getvfs(&fh.fh_fsid)) == NULL) {
+		error = ESTALE;
+		goto bad;
 	}
-	vrele(vp);
-	return (error);
-}
+
+	if ((error = VFS_FHTOVP(mp, &fh.fh_fid, &vp)) != 0) {
+		vp = NULL;	/* most likely unnecessary sanity for bad: */
+		goto bad;
+	}
+
+	/* Now do an effective vn_open */
+
+	if (vp->v_type == VSOCK) {
+		error = EOPNOTSUPP;
+		goto bad;
+	}
+	if (flags & FREAD) {
+		if ((error = VOP_ACCESS(vp, VREAD, cred, p)) != 0)
+			goto bad;
+	}
+	if (flags & (FWRITE | O_TRUNC)) {
+		if (vp->v_type == VDIR) {
+			error = EISDIR;
+			goto bad;
+		}
+		if ((error = vn_writechk(vp)) != 0 ||
+		    (error = VOP_ACCESS(vp, VWRITE, cred, p)) != 0)
+			goto bad;
+	}
+	if (flags & O_TRUNC) {
+		VOP_UNLOCK(vp, 0, p);				/* XXX */
+		VOP_LEASE(vp, p, cred, LEASE_WRITE);
+		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);	/* XXX */
+		VATTR_NULL(&va);
+		va.va_size = 0;
+		if ((error = VOP_SETATTR(vp, &va, cred, p)) != 0)
+			goto bad;
+	}
+	if ((error = VOP_OPEN(vp, flags, cred, p)) != 0)
+		goto bad;
+	if (flags & FWRITE)
+		vp->v_writecount++;
+
+	/* done with modified vn_open, now finish what sys_open does. */
+
+	fp->f_flag = flags & FMASK;
+	fp->f_type = DTYPE_VNODE;
+	fp->f_ops = &vnops;
+	fp->f_data = (caddr_t)vp;
+	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;
+		VOP_UNLOCK(vp, 0, p);
+		error = VOP_ADVLOCK(vp, (caddr_t)fp, F_SETLK, &lf, type);
+		if (error) {
+			/* closef will vn_close the file for us. */
+			fdremove(fdp, indx);
+			closef(fp, p);
+			return (error);
+		}
+		vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+		fp->f_flag |= FHASLOCK;
+	}
+	VOP_UNLOCK(vp, 0, p);
+	*retval = indx;
+	FILE_SET_MATURE(fp);
+	return (0);
+
+bad:
+	fdremove(fdp, indx);
+	closef(fp, p);
+	if (vp != NULL)
+		vput(vp);
+	return (error);
+}
+
+/* ARGSUSED */
+int
+sys_fhstat(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_fhstat_args /* {
+		syscallarg(const fhandle_t *) fhp;
+		syscallarg(struct stat *) sb;
+	} */ *uap = v;
+	struct stat sb;
+	int error;
+	fhandle_t fh;
+	struct mount *mp;
+	struct vnode *vp;
+
+	/*
+	 * Must be super user
+	 */
+	if ((error = suser(p->p_ucred, &p->p_acflag)))
+		return (error);
+
+	if ((error = copyin(SCARG(uap, 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);
+	error = vn_stat(vp, &sb, p);
+	vput(vp);
+	if (error)
+		return (error);
+	error = copyout(&sb, SCARG(uap, sb), sizeof(sb));
+	return (error);
+}
+
+/* ARGSUSED */
+int
+sys_fhstatfs(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_fhstatfs_args /*
+		syscallarg(const fhandle_t *) fhp;
+		syscallarg(struct statfs *) buf;
+	} */ *uap = v;
+	struct statfs sp;
+	fhandle_t fh;
+	struct mount *mp;
+	struct vnode *vp;
+	int error;
+
+	/*
+	 * Must be super user
+	 */
+	if ((error = suser(p->p_ucred, &p->p_acflag)))
+		return (error);
+
+	if ((error = copyin(SCARG(uap, 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;
+	vput(vp);
+	if ((error = VFS_STATFS(mp, &sp, p)) != 0)
+		return (error);
+	sp.f_flags = mp->mnt_flag & MNT_VISFLAGMASK;
+	return (copyout(&sp, SCARG(uap, buf), sizeof(sp)));
+}
+
+/*
+ * Create a special file.
+ */
+/* ARGSUSED */
+int
+sys_mknod(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_mknod_args /* {
+		syscallarg(char *) path;
+		syscallarg(int) mode;
+		syscallarg(int) dev;
+	} */ *uap = v;
+	register struct vnode *vp;
+	struct vattr vattr;
+	int error;
+	int whiteout = 0;
+	struct nameidata nd;
+
+	if ((error = suser(p->p_ucred, &p->p_acflag)) != 0)
+		return (error);
+	if (p->p_fd->fd_rdir)
+		return (EINVAL);
+	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
+		return (error);
+	vp = nd.ni_vp;
+	if (vp != NULL)
+		error = EEXIST;
+	else {
+		VATTR_NULL(&vattr);
+		vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ p->p_fd->fd_cmask;
+		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 (!error) {
+		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+		if (whiteout) {
+			error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, CREATE);
+			if (error)
+				VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
+			vput(nd.ni_dvp);
+		} else {
+			error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp,
+						&nd.ni_cnd, &vattr);
+		}
+	} else {
+		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
+		if (nd.ni_dvp == vp)
+			vrele(nd.ni_dvp);
+		else
+			vput(nd.ni_dvp);
+		if (vp)
+			vrele(vp);
+	}
+	return (error);
+}
+
+/*
+ * Create a named pipe.
+ */
+/* ARGSUSED */
+int
+sys_mkfifo(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+#ifndef FIFO
+	return (EOPNOTSUPP);
+#else
+	register struct sys_mkfifo_args /* {
+		syscallarg(char *) path;
+		syscallarg(int) mode;
+	} */ *uap = v;
+	struct vattr vattr;
+	int error;
+	struct nameidata nd;
+
+	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
+		return (error);
+	if (nd.ni_vp != NULL) {
+		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
+		if (nd.ni_dvp == nd.ni_vp)
+			vrele(nd.ni_dvp);
+		else
+			vput(nd.ni_dvp);
+		vrele(nd.ni_vp);
+		return (EEXIST);
+	}
+	VATTR_NULL(&vattr);
+	vattr.va_type = VFIFO;
+	vattr.va_mode = (SCARG(uap, mode) & ALLPERMS) &~ p->p_fd->fd_cmask;
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	return (VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr));
+#endif /* FIFO */
+}
+
+/*
+ * Make a hard file link.
+ */
+/* ARGSUSED */
+int
+sys_link(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_link_args /* {
+		syscallarg(char *) path;
+		syscallarg(char *) link;
+	} */ *uap = v;
+	register struct vnode *vp;
+	struct nameidata nd;
+	int error;
+	int flags;
+
+	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
+		return (error);
+	vp = nd.ni_vp;
+
+	flags = LOCKPARENT;
+	if (vp->v_type == VDIR) {
+		flags |= STRIPSLASHES;
+	}
+
+	NDINIT(&nd, CREATE, flags, UIO_USERSPACE, SCARG(uap, link), p);
+	if ((error = namei(&nd)) != 0)
+		goto out;
+	if (nd.ni_vp) {
+		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
+		if (nd.ni_dvp == nd.ni_vp)
+			vrele(nd.ni_dvp);
+		else
+			vput(nd.ni_dvp);
+		vrele(nd.ni_vp);
+		error = EEXIST;
+		goto out;
+	}
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	error = VOP_LINK(nd.ni_dvp, vp, &nd.ni_cnd);
+out:
+	vrele(vp);
+	return (error);
+}
 
 /*
  * Make a symbolic link.
  */
 /* ARGSUSED */
+int
 sys_symlink(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -966,10 +1394,11 @@
 	struct nameidata nd;
 
 	MALLOC(path, char *, MAXPATHLEN, M_NAMEI, M_WAITOK);
-	if (error = copyinstr(SCARG(uap, path), path, MAXPATHLEN, (size_t *)0))
+	error = copyinstr(SCARG(uap, path), path, MAXPATHLEN, NULL);
+	if (error)
 		goto out;
 	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, link), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		goto out;
 	if (nd.ni_vp) {
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
@@ -994,6 +1423,7 @@
  * Delete a whiteout from the filesystem.
  */
 /* ARGSUSED */
+int
 sys_undelete(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1023,7 +1453,7 @@
 	}
 
 	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-	if (error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, DELETE))
+	if ((error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, DELETE)) != 0)
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
 	vput(nd.ni_dvp);
 	return (error);
@@ -1033,6 +1463,7 @@
  * Delete a name from the filesystem.
  */
 /* ARGSUSED */
+int
 sys_unlink(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1045,42 +1476,39 @@
 	int error;
 	struct nameidata nd;
 
-	NDINIT(&nd, DELETE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	NDINIT(&nd, DELETE, LOCKPARENT | LOCKLEAF, UIO_USERSPACE,
+	    SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
-	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
-
-	if (vp->v_type != VDIR ||
-	    (error = suser(p->p_ucred, &p->p_acflag)) == 0) {
-		/*
-		 * The root of a mounted filesystem cannot be deleted.
-		 */
-		if (vp->v_flag & VROOT)
-			error = EBUSY;
-		else
-			(void)vnode_pager_uncache(vp);
-	}
 
-	if (!error) {
-		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-		error = VOP_REMOVE(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd);
-	} else {
+	/*
+	 * The root of a mounted filesystem cannot be deleted.
+	 */
+	if (vp->v_flag & VROOT) {
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
 		if (nd.ni_dvp == vp)
 			vrele(nd.ni_dvp);
 		else
 			vput(nd.ni_dvp);
-		if (vp != NULLVP)
-			vput(vp);
+		vput(vp);
+		error = EBUSY;
+		goto out;
 	}
+
+	(void)uvm_vnp_uncache(vp);
+
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	error = VOP_REMOVE(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd);
+out:
 	return (error);
 }
 
 /*
  * Reposition read/write file offset.
  */
+int
 sys_lseek(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1096,24 +1524,38 @@
 	register struct filedesc *fdp = p->p_fd;
 	register struct file *fp;
 	struct vattr vattr;
-	int error;
+	struct vnode *vp;
+	int error, special;
 
-	if ((u_int)SCARG(uap, fd) >= fdp->fd_nfiles ||
-	    (fp = fdp->fd_ofiles[SCARG(uap, fd)]) == NULL)
+	if ((fp = fd_getfile(fdp, SCARG(uap, fd))) == NULL)
 		return (EBADF);
 	if (fp->f_type != DTYPE_VNODE)
 		return (ESPIPE);
+	vp = (struct vnode *)fp->f_data;
+	if (vp->v_type == VFIFO)
+		return (ESPIPE);
+	if (vp->v_type == VCHR)
+		special = 1;
+	else
+		special = 0;
 	switch (SCARG(uap, whence)) {
-	case L_INCR:
+	case SEEK_CUR:
+		if (!special && fp->f_offset + SCARG(uap, offset) < 0)
+			return (EINVAL);
 		fp->f_offset += SCARG(uap, offset);
 		break;
-	case L_XTND:
-		if (error =
-		    VOP_GETATTR((struct vnode *)fp->f_data, &vattr, cred, p))
+	case SEEK_END:
+		error = VOP_GETATTR((struct vnode *)fp->f_data, &vattr,
+				    cred, p);
+		if (error)
 			return (error);
+		if (!special && (off_t)vattr.va_size + SCARG(uap, offset) < 0)
+			return (EINVAL);
 		fp->f_offset = SCARG(uap, offset) + vattr.va_size;
 		break;
-	case L_SET:
+	case SEEK_SET:
+		if (!special && SCARG(uap, offset) < 0)
+			return (EINVAL);
 		fp->f_offset = SCARG(uap, offset);
 		break;
 	default:
@@ -1126,6 +1568,7 @@
 /*
  * Check access permissions.
  */
+int
 sys_access(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1140,13 +1583,15 @@
 	int error, flags, t_gid, t_uid;
 	struct nameidata nd;
 
+	if (SCARG(uap, flags) & ~(R_OK | W_OK | X_OK))
+		return (EINVAL);
 	t_uid = cred->cr_uid;
 	t_gid = cred->cr_gid;
 	cred->cr_uid = p->p_cred->p_ruid;
 	cred->cr_gid = p->p_cred->p_rgid;
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		goto out1;
 	vp = nd.ni_vp;
 
@@ -1173,6 +1618,7 @@
  * Get file status; this version follows links.
  */
 /* ARGSUSED */
+int
 sys_stat(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1188,12 +1634,15 @@
 
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	error = vn_stat(nd.ni_vp, &sb, p);
 	vput(nd.ni_vp);
 	if (error)
 		return (error);
+	/* Don't let non-root see generation numbers (for NFS security) */
+	if (suser(p->p_ucred, &p->p_acflag))
+		sb.st_gen = 0;
 	error = copyout((caddr_t)&sb, (caddr_t)SCARG(uap, ub), sizeof (sb));
 	return (error);
 }
@@ -1202,6 +1651,7 @@
  * Get file status; this version does not follow links.
  */
 /* ARGSUSED */
+int
 sys_lstat(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1211,47 +1661,21 @@
 		syscallarg(char *) path;
 		syscallarg(struct stat *) ub;
 	} */ *uap = v;
+	struct stat sb;
 	int error;
-	struct vnode *vp, *dvp;
-	struct stat sb, sb1;
 	struct nameidata nd;
 
-	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | LOCKPARENT, UIO_USERSPACE,
+	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
-	/*
-	 * For symbolic links, always return the attributes of its
-	 * containing directory, except for mode, size, and links.
-	 */
-	vp = nd.ni_vp;
-	dvp = nd.ni_dvp;
-	if (vp->v_type != VLNK) {
-		if (dvp == vp)
-			vrele(dvp);
-		else
-			vput(dvp);
-		error = vn_stat(vp, &sb, p);
-		vput(vp);
-		if (error)
-			return (error);
-	} else {
-		error = vn_stat(dvp, &sb, p);
-		vput(dvp);
-		if (error) {
-			vput(vp);
-			return (error);
-		}
-		error = vn_stat(vp, &sb1, p);
-		vput(vp);
-		if (error)
-			return (error);
-		sb.st_mode &= ~S_IFDIR;
-		sb.st_mode |= S_IFLNK;
-		sb.st_nlink = sb1.st_nlink;
-		sb.st_size = sb1.st_size;
-		sb.st_blocks = sb1.st_blocks;
-	}
+	error = vn_stat(nd.ni_vp, &sb, p);
+	vput(nd.ni_vp);
+	if (error)
+		return (error);
+	/* Don't let non-root see generation numbers (for NFS security) */
+	if (suser(p->p_ucred, &p->p_acflag))
+		sb.st_gen = 0;
 	error = copyout((caddr_t)&sb, (caddr_t)SCARG(uap, ub), sizeof (sb));
 	return (error);
 }
@@ -1260,6 +1684,7 @@
  * Get configurable pathname variables.
  */
 /* ARGSUSED */
+int
 sys_pathconf(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1274,7 +1699,7 @@
 
 	NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	error = VOP_PATHCONF(nd.ni_vp, SCARG(uap, name), retval);
 	vput(nd.ni_vp);
@@ -1285,6 +1710,7 @@
  * Return target name of a symbolic link.
  */
 /* ARGSUSED */
+int
 sys_readlink(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1293,7 +1719,7 @@
 	register struct sys_readlink_args /* {
 		syscallarg(char *) path;
 		syscallarg(char *) buf;
-		syscallarg(int) count;
+		syscallarg(size_t) count;
 	} */ *uap = v;
 	register struct vnode *vp;
 	struct iovec aiov;
@@ -1303,7 +1729,7 @@
 
 	NDINIT(&nd, LOOKUP, NOFOLLOW | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	if (vp->v_type != VLNK)
@@ -1329,6 +1755,7 @@
  * Change flags of a file given a path name.
  */
 /* ARGSUSED */
+int
 sys_chflags(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1336,7 +1763,7 @@
 {
 	register struct sys_chflags_args /* {
 		syscallarg(char *) path;
-		syscallarg(int) flags;
+		syscallarg(unsigned int) flags;
 	} */ *uap = v;
 	register struct vnode *vp;
 	struct vattr vattr;
@@ -1344,18 +1771,29 @@
 	struct nameidata nd;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
+	else if (SCARG(uap, flags) == VNOVAL)
+		error = EINVAL;
 	else {
+		if (suser(p->p_ucred, &p->p_acflag)) {
+			if ((error = VOP_GETATTR(vp, &vattr, p->p_ucred, p)) != 0)
+				goto out;
+			if (vattr.va_type == VCHR || vattr.va_type == VBLK) {
+				error = EINVAL;
+				goto out;
+			}
+		}
 		VATTR_NULL(&vattr);
 		vattr.va_flags = SCARG(uap, flags);
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
+out:
 	vput(vp);
 	return (error);
 }
@@ -1364,33 +1802,47 @@
  * Change flags of a file given a file descriptor.
  */
 /* ARGSUSED */
+int
 sys_fchflags(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_fchflags_args /* {
+	struct sys_fchflags_args /* {
 		syscallarg(int) fd;
-		syscallarg(int) flags;
+		syscallarg(unsigned int) flags;
 	} */ *uap = v;
 	struct vattr vattr;
 	struct vnode *vp;
 	struct file *fp;
 	int error;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	vp = (struct vnode *)fp->f_data;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
+	else if (SCARG(uap, flags) == VNOVAL)
+		error = EINVAL;
 	else {
+		if (suser(p->p_ucred, &p->p_acflag)) {
+			if ((error = VOP_GETATTR(vp, &vattr, p->p_ucred, p))
+			    != 0)
+				goto out;
+			if (vattr.va_type == VCHR || vattr.va_type == VBLK) {
+				error = EINVAL;
+				goto out;
+			}
+		}
 		VATTR_NULL(&vattr);
 		vattr.va_flags = SCARG(uap, flags);
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
-	VOP_UNLOCK(vp);
+out:
+	VOP_UNLOCK(vp, 0, p);
+	FRELE(fp);
 	return (error);
 }
 
@@ -1398,6 +1850,7 @@
  * Change mode of a file given path name.
  */
 /* ARGSUSED */
+int
 sys_chmod(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1412,12 +1865,15 @@
 	int error;
 	struct nameidata nd;
 
+	if (SCARG(uap, mode) & ~(S_IFMT | ALLPERMS))
+		return (EINVAL);
+
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
 	else {
@@ -1433,12 +1889,13 @@
  * Change mode of a file given a file descriptor.
  */
 /* ARGSUSED */
+int
 sys_fchmod(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_fchmod_args /* {
+	struct sys_fchmod_args /* {
 		syscallarg(int) fd;
 		syscallarg(int) mode;
 	} */ *uap = v;
@@ -1447,11 +1904,14 @@
 	struct file *fp;
 	int error;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if (SCARG(uap, mode) & ~(S_IFMT | ALLPERMS))
+		return (EINVAL);
+
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	vp = (struct vnode *)fp->f_data;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
 	else {
@@ -1459,7 +1919,8 @@
 		vattr.va_mode = SCARG(uap, mode) & ALLPERMS;
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
-	VOP_UNLOCK(vp);
+	VOP_UNLOCK(vp, 0, p);
+	FRELE(fp);
 	return (error);
 }
 
@@ -1467,6 +1928,7 @@
  * Set ownership given a path name.
  */
 /* ARGSUSED */
+int
 sys_chown(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1481,21 +1943,87 @@
 	struct vattr vattr;
 	int error;
 	struct nameidata nd;
+	u_short mode;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
+		return (error);
+	vp = nd.ni_vp;
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+	if (vp->v_mount->mnt_flag & MNT_RDONLY)
+		error = EROFS;
+	else {
+		if ((SCARG(uap, uid) != -1 || SCARG(uap, gid) != -1) &&
+		    (suser(p->p_ucred, &p->p_acflag) || suid_clear)) {
+			error = VOP_GETATTR(vp, &vattr, p->p_ucred, p);
+			if (error)
+				goto out;
+			mode = vattr.va_mode & ~(VSUID | VSGID);
+			if (mode == vattr.va_mode)
+				mode = VNOVAL;
+		}
+		else
+			mode = VNOVAL;
+		VATTR_NULL(&vattr);
+		vattr.va_uid = SCARG(uap, uid);
+		vattr.va_gid = SCARG(uap, gid);
+		vattr.va_mode = mode;
+		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
+	}
+out:
+	vput(vp);
+	return (error);
+}
+
+/*
+ * Set ownership given a path name, without following links.
+ */
+/* ARGSUSED */
+int
+sys_lchown(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_lchown_args /* {
+		syscallarg(char *) path;
+		syscallarg(int) uid;
+		syscallarg(int) gid;
+	} */ *uap = v;
+	register struct vnode *vp;
+	struct vattr vattr;
+	int error;
+	struct nameidata nd;
+	u_short mode;
+
+	NDINIT(&nd, LOOKUP, NOFOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
 	else {
+		if ((SCARG(uap, uid) != -1 || SCARG(uap, gid) != -1) &&
+		    (suser(p->p_ucred, &p->p_acflag) || suid_clear)) {
+			error = VOP_GETATTR(vp, &vattr, p->p_ucred, p);
+			if (error)
+				goto out;
+			mode = vattr.va_mode & ~(VSUID | VSGID);
+			if (mode == vattr.va_mode)
+				mode = VNOVAL;
+		}
+		else
+			mode = VNOVAL;
 		VATTR_NULL(&vattr);
 		vattr.va_uid = SCARG(uap, uid);
 		vattr.va_gid = SCARG(uap, gid);
+		vattr.va_mode = mode;
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
+out:
 	vput(vp);
 	return (error);
 }
@@ -1504,42 +2032,58 @@
  * Set ownership given a file descriptor.
  */
 /* ARGSUSED */
+int
 sys_fchown(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_fchown_args /* {
+	struct sys_fchown_args /* {
 		syscallarg(int) fd;
 		syscallarg(int) uid;
 		syscallarg(int) gid;
 	} */ *uap = v;
-	struct vattr vattr;
 	struct vnode *vp;
-	struct file *fp;
+	struct vattr vattr;
 	int error;
+	struct file *fp;
+	u_short mode;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	vp = (struct vnode *)fp->f_data;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
 	else {
+		if ((SCARG(uap, uid) != -1 || SCARG(uap, gid) != -1) &&
+		    (suser(p->p_ucred, &p->p_acflag) || suid_clear)) {
+			error = VOP_GETATTR(vp, &vattr, p->p_ucred, p);
+			if (error)
+				goto out;
+			mode = vattr.va_mode & ~(VSUID | VSGID);
+			if (mode == vattr.va_mode)
+				mode = VNOVAL;
+		} else
+			mode = VNOVAL;
 		VATTR_NULL(&vattr);
 		vattr.va_uid = SCARG(uap, uid);
 		vattr.va_gid = SCARG(uap, gid);
+		vattr.va_mode = mode;
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
-	VOP_UNLOCK(vp);
+out:
+	VOP_UNLOCK(vp, 0, p);
+	FRELE(fp);
 	return (error);
 }
 
 /*
- * Set the access and modification times of a file.
+ * Set the access and modification times given a path name.
  */
 /* ARGSUSED */
+int
 sys_utimes(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1560,32 +2104,97 @@
 		microtime(&tv[0]);
 		tv[1] = tv[0];
 		vattr.va_vaflags |= VA_UTIMES_NULL;
-	} else if (error = copyin((caddr_t)SCARG(uap, tptr), (caddr_t)tv,
-	    sizeof (tv)))
-  		return (error);
+	} else {
+		error = copyin((caddr_t)SCARG(uap, tptr), (caddr_t)tv,
+		    sizeof (tv));
+		if (error)
+			return (error);
+		/* XXX workaround timeval matching the VFS constant VNOVAL */
+		if (tv[0].tv_sec == VNOVAL)
+			tv[0].tv_sec = VNOVAL - 1;
+		if (tv[1].tv_sec == VNOVAL)
+			tv[1].tv_sec = VNOVAL - 1;
+	}
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
 	else {
-		vattr.va_atime.ts_sec = tv[0].tv_sec;
-		vattr.va_atime.ts_nsec = tv[0].tv_usec * 1000;
-		vattr.va_mtime.ts_sec = tv[1].tv_sec;
-		vattr.va_mtime.ts_nsec = tv[1].tv_usec * 1000;
+		vattr.va_atime.tv_sec = tv[0].tv_sec;
+		vattr.va_atime.tv_nsec = tv[0].tv_usec * 1000;
+		vattr.va_mtime.tv_sec = tv[1].tv_sec;
+		vattr.va_mtime.tv_nsec = tv[1].tv_usec * 1000;
 		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
 	}
 	vput(vp);
 	return (error);
 }
 
+
+/*
+ * Set the access and modification times given a file descriptor.
+ */
+/* ARGSUSED */
+int
+sys_futimes(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	register struct sys_futimes_args /* {
+		syscallarg(int) fd;
+		syscallarg(struct timeval *) tptr;
+	} */ *uap = v;
+	struct vnode *vp;
+	struct timeval tv[2];
+	struct vattr vattr;
+	int error;
+	struct file *fp;
+
+	VATTR_NULL(&vattr);
+	if (SCARG(uap, tptr) == NULL) {
+		microtime(&tv[0]);
+		tv[1] = tv[0];
+		vattr.va_vaflags |= VA_UTIMES_NULL;
+	} else {
+		error = copyin((caddr_t)SCARG(uap, tptr), (caddr_t)tv,
+		    sizeof (tv));
+		if (error)
+			return (error);
+		/* XXX workaround timeval matching the VFS constant VNOVAL */
+		if (tv[0].tv_sec == VNOVAL)
+			tv[0].tv_sec = VNOVAL - 1;
+		if (tv[1].tv_sec == VNOVAL)
+			tv[1].tv_sec = VNOVAL - 1;
+	}
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
+		return (error);
+	vp = (struct vnode *)fp->f_data;
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+	if (vp->v_mount->mnt_flag & MNT_RDONLY)
+		error = EROFS;
+	else {
+		vattr.va_atime.tv_sec = tv[0].tv_sec;
+		vattr.va_atime.tv_nsec = tv[0].tv_usec * 1000;
+		vattr.va_mtime.tv_sec = tv[1].tv_sec;
+		vattr.va_mtime.tv_nsec = tv[1].tv_usec * 1000;
+		error = VOP_SETATTR(vp, &vattr, p->p_ucred, p);
+	}
+	VOP_UNLOCK(vp, 0, p);
+	FRELE(fp);
+	return (error);
+}
+
 /*
  * Truncate a file given its path name.
  */
 /* ARGSUSED */
+int
 sys_truncate(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1602,11 +2211,11 @@
 	struct nameidata nd;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_type == VDIR)
 		error = EISDIR;
 	else if ((error = vn_writechk(vp)) == 0 &&
@@ -1623,12 +2232,13 @@
  * Truncate a file given a file descriptor.
  */
 /* ARGSUSED */
+int
 sys_ftruncate(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_ftruncate_args /* {
+	struct sys_ftruncate_args /* {
 		syscallarg(int) fd;
 		syscallarg(int) pad;
 		syscallarg(off_t) length;
@@ -1638,13 +2248,15 @@
 	struct file *fp;
 	int error;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
-	if ((fp->f_flag & FWRITE) == 0)
-		return (EINVAL);
+	if ((fp->f_flag & FWRITE) == 0) {
+		error = EINVAL;
+		goto bad;
+	}
 	vp = (struct vnode *)fp->f_data;
 	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	if (vp->v_type == VDIR)
 		error = EISDIR;
 	else if ((error = vn_writechk(vp)) == 0) {
@@ -1652,7 +2264,9 @@
 		vattr.va_size = SCARG(uap, length);
 		error = VOP_SETATTR(vp, &vattr, fp->f_cred, p);
 	}
-	VOP_UNLOCK(vp);
+	VOP_UNLOCK(vp, 0, p);
+bad:
+	FRELE(fp);
 	return (error);
 }
 
@@ -1660,6 +2274,7 @@
  * Sync an open file.
  */
 /* ARGSUSED */
+int
 sys_fsync(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1668,16 +2283,22 @@
 	struct sys_fsync_args /* {
 		syscallarg(int) fd;
 	} */ *uap = v;
-	register struct vnode *vp;
+	struct vnode *vp;
 	struct file *fp;
 	int error;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
 		return (error);
 	vp = (struct vnode *)fp->f_data;
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	error = VOP_FSYNC(vp, fp->f_cred, MNT_WAIT, p);
-	VOP_UNLOCK(vp);
+#ifdef FFS_SOFTUPDATES
+	if (error == 0 && vp->v_mount && (vp->v_mount->mnt_flag & MNT_SOFTDEP))
+		error = softdep_fsync(vp);
+#endif
+
+	VOP_UNLOCK(vp, 0, p);
+	FRELE(fp);
 	return (error);
 }
 
@@ -1686,6 +2307,7 @@
  * or both not be directories.  If target is a directory, it must be empty.
  */
 /* ARGSUSED */
+int
 sys_rename(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1698,15 +2320,24 @@
 	register struct vnode *tvp, *fvp, *tdvp;
 	struct nameidata fromnd, tond;
 	int error;
+	int flags;
 
 	NDINIT(&fromnd, DELETE, WANTPARENT | SAVESTART, UIO_USERSPACE,
 	    SCARG(uap, from), p);
-	if (error = namei(&fromnd))
+	if ((error = namei(&fromnd)) != 0)
 		return (error);
 	fvp = fromnd.ni_vp;
-	NDINIT(&tond, RENAME, LOCKPARENT | LOCKLEAF | NOCACHE | SAVESTART,
+
+	flags = LOCKPARENT | LOCKLEAF | NOCACHE | SAVESTART;
+	/*
+	 * rename("foo/", "bar/");  is  OK
+	 */
+	if (fvp->v_type == VDIR)
+		flags |= STRIPSLASHES;
+
+	NDINIT(&tond, RENAME, flags,
 	    UIO_USERSPACE, SCARG(uap, to), p);
-	if (error = namei(&tond)) {
+	if ((error = namei(&tond)) != 0) {
 		VOP_ABORTOP(fromnd.ni_dvp, &fromnd.ni_cnd);
 		vrele(fromnd.ni_dvp);
 		vrele(fvp);
@@ -1727,21 +2358,19 @@
 		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.
+	 * same inode number)
 	 */
-	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))
+	if (fvp == tvp)
 		error = -1;
 out:
 	if (!error) {
 		VOP_LEASE(tdvp, p, p->p_ucred, LEASE_WRITE);
 		if (fromnd.ni_dvp != tdvp)
 			VOP_LEASE(fromnd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-		if (tvp)
+		if (tvp) {
+			(void)uvm_vnp_uncache(tvp);
 			VOP_LEASE(tvp, p, p->p_ucred, LEASE_WRITE);
+		}
 		error = VOP_RENAME(fromnd.ni_dvp, fromnd.ni_vp, &fromnd.ni_cnd,
 				   tond.ni_dvp, tond.ni_vp, &tond.ni_cnd);
 	} else {
@@ -1771,6 +2400,7 @@
  * Make a directory file.
  */
 /* ARGSUSED */
+int
 sys_mkdir(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1785,8 +2415,9 @@
 	int error;
 	struct nameidata nd;
 
-	NDINIT(&nd, CREATE, LOCKPARENT, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	NDINIT(&nd, CREATE, LOCKPARENT | STRIPSLASHES,
+	    UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	if (vp != NULL) {
@@ -1812,6 +2443,7 @@
  * Remove a directory file.
  */
 /* ARGSUSED */
+int
 sys_rmdir(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -1826,7 +2458,7 @@
 
 	NDINIT(&nd, DELETE, LOCKPARENT | LOCKLEAF, UIO_USERSPACE,
 	    SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
 	if (vp->v_type != VDIR) {
@@ -1837,7 +2469,7 @@
 	 * No rmdir "." please.
 	 */
 	if (nd.ni_dvp == vp) {
-		error = EINVAL;
+		error = EBUSY;
 		goto out;
 	}
 	/*
@@ -1864,32 +2496,39 @@
 /*
  * Read a block of directory entries in a file system independent format.
  */
+int
 sys_getdirentries(p, v, retval)
 	struct proc *p;
 	void *v;
 	register_t *retval;
 {
-	register struct sys_getdirentries_args /* {
+	struct sys_getdirentries_args /* {
 		syscallarg(int) fd;
 		syscallarg(char *) buf;
-		syscallarg(u_int) count;
+		syscallarg(int) count;
 		syscallarg(long *) basep;
 	} */ *uap = v;
-	register struct vnode *vp;
+	struct vnode *vp;
 	struct file *fp;
 	struct uio auio;
 	struct iovec aiov;
 	long loff;
 	int error, eofflag;
 
-	if (error = getvnode(p->p_fd, SCARG(uap, fd), &fp))
-		return (error);
-	if ((fp->f_flag & FREAD) == 0)
-		return (EBADF);
+	if (SCARG(uap, count) < 0)
+		return EINVAL;
+	if ((error = getvnode(p->p_fd, SCARG(uap, fd), &fp)) != 0)
+		return (error);
+	if ((fp->f_flag & FREAD) == 0) {
+		error = EBADF;
+		goto bad;
+	}
 	vp = (struct vnode *)fp->f_data;
 unionread:
-	if (vp->v_type != VDIR)
-		return (EINVAL);
+	if (vp->v_type != VDIR) {
+		error = EINVAL;
+		goto bad;
+	}
 	aiov.iov_base = SCARG(uap, buf);
 	aiov.iov_len = SCARG(uap, count);
 	auio.uio_iov = &aiov;
@@ -1898,57 +2537,19 @@
 	auio.uio_segflg = UIO_USERSPACE;
 	auio.uio_procp = p;
 	auio.uio_resid = SCARG(uap, count);
-	VOP_LOCK(vp);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
 	loff = auio.uio_offset = fp->f_offset;
-	error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, (u_long *)0, 0);
+	error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, 0, 0);
 	fp->f_offset = auio.uio_offset;
-	VOP_UNLOCK(vp);
+	VOP_UNLOCK(vp, 0, p);
 	if (error)
-		return (error);
-
-#ifdef UNION
-{
-	extern int (**union_vnodeop_p)();
-	extern struct vnode *union_dircache __P((struct vnode *));
-
+		goto bad;
 	if ((SCARG(uap, count) == auio.uio_resid) &&
-	    (vp->v_op == union_vnodeop_p)) {
-		struct vnode *lvp;
-
-		lvp = union_dircache(vp);
-		if (lvp != NULLVP) {
-			struct vattr va;
-
-			/*
-			 * If the directory is opaque,
-			 * then don't show lower entries
-			 */
-			error = VOP_GETATTR(vp, &va, fp->f_cred, p);
-			if (va.va_flags & OPAQUE) {
-				vput(lvp);
-				lvp = NULL;
-			}
-		}
-		
-		if (lvp != NULLVP) {
-			error = VOP_OPEN(lvp, FREAD, fp->f_cred, p);
-			VOP_UNLOCK(lvp);
-
-			if (error) {
-				vrele(lvp);
-				return (error);
-			}
-			fp->f_data = (caddr_t) lvp;
-			fp->f_offset = 0;
-			error = vn_close(vp, FREAD, fp->f_cred, p);
-			if (error)
-				return (error);
-			vp = lvp;
-			goto unionread;
-		}
-	}
-}
-#endif /* UNION */
+	    union_check_p &&
+	    (union_check_p(p, &vp, fp, auio, &error) != 0))
+		goto unionread;
+	if (error)
+		goto bad;
 
 	if ((SCARG(uap, count) == auio.uio_resid) &&
 	    (vp->v_flag & VROOT) &&
@@ -1964,6 +2565,8 @@
 	error = copyout((caddr_t)&loff, (caddr_t)SCARG(uap, basep),
 	    sizeof(long));
 	*retval = SCARG(uap, count) - auio.uio_resid;
+bad:
+	FRELE(fp);
 	return (error);
 }
 
@@ -1983,7 +2586,7 @@
 
 	fdp = p->p_fd;
 	*retval = fdp->fd_cmask;
-	fdp->fd_cmask = SCARG(uap, newmask) & ALLPERMS;
+	fdp->fd_cmask = SCARG(uap, newmask) & ACCESSPERMS;
 	return (0);
 }
 
@@ -1992,6 +2595,7 @@
  * away from vnode.
  */
 /* ARGSUSED */
+int
 sys_revoke(p, v, retval)
 	struct proc *p;
 	void *v;
@@ -2006,20 +2610,16 @@
 	struct nameidata nd;
 
 	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
-	if (error = namei(&nd))
+	if ((error = namei(&nd)) != 0)
 		return (error);
 	vp = nd.ni_vp;
-	if (vp->v_type != VCHR && vp->v_type != VBLK) {
-		error = EINVAL;
-		goto out;
-	}
-	if (error = VOP_GETATTR(vp, &vattr, p->p_ucred, p))
+	if ((error = VOP_GETATTR(vp, &vattr, p->p_ucred, p)) != 0)
 		goto out;
 	if (p->p_ucred->cr_uid != vattr.va_uid &&
 	    (error = suser(p->p_ucred, &p->p_acflag)))
 		goto out;
-	if (vp->v_usecount > 1 || (vp->v_flag & VALIASED))
-		vgoneall(vp);
+	if (vp->v_usecount > 1 || (vp->v_flag & (VALIASED | VLAYER)))
+		VOP_REVOKE(vp, REVOKEALL);
 out:
 	vrele(vp);
 	return (error);
@@ -2027,7 +2627,10 @@
 
 /*
  * Convert a user file descriptor to a kernel file entry.
+ *
+ * On return *fpp is FREF:ed.
  */
+int
 getvnode(fdp, fd, fpp)
 	struct filedesc *fdp;
 	struct file **fpp;
@@ -2035,11 +2638,579 @@
 {
 	struct file *fp;
 
-	if ((u_int)fd >= fdp->fd_nfiles ||
-	    (fp = fdp->fd_ofiles[fd]) == NULL)
+	if ((fp = fd_getfile(fdp, fd)) == NULL)
 		return (EBADF);
 	if (fp->f_type != DTYPE_VNODE)
 		return (EINVAL);
+	FREF(fp);
 	*fpp = fp;
+
 	return (0);
 }
+
+/*
+ * Positional read system call.
+ */
+int
+sys_pread(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_pread_args /* {
+		syscallarg(int) fd;
+		syscallarg(void *) buf;
+		syscallarg(size_t) nbyte;
+		syscallarg(int) pad;
+		syscallarg(off_t) offset;
+	} */ *uap = v;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp;
+	off_t offset;
+	int fd = SCARG(uap, fd);
+
+	if ((fp = fd_getfile(fdp, fd)) == NULL)
+		return (EBADF);
+	if ((fp->f_flag & FREAD) == 0)
+		return (EBADF);
+
+	vp = (struct vnode *)fp->f_data;
+	if (fp->f_type != DTYPE_VNODE || vp->v_type == VFIFO) {
+		return (ESPIPE);
+	}
+
+	offset = SCARG(uap, offset);
+
+	FREF(fp);
+
+	/* dofileread() will FRELE the descriptor for us */
+	return (dofileread(p, fd, fp, SCARG(uap, buf), SCARG(uap, nbyte),
+	    &offset, retval));
+}
+
+/*
+ * Positional scatter read system call.
+ */
+int
+sys_preadv(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_preadv_args /* {
+		syscallarg(int) fd;
+		syscallarg(const struct iovec *) iovp;
+		syscallarg(int) iovcnt;
+		syscallarg(int) pad;
+		syscallarg(off_t) offset;
+	} */ *uap = v;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp;
+	off_t offset;
+	int fd = SCARG(uap, fd);
+
+	if ((fp = fd_getfile(fdp, fd)) == NULL)
+		return (EBADF);
+	if ((fp->f_flag & FREAD) == 0)
+		return (EBADF);
+
+	vp = (struct vnode *)fp->f_data;
+	if (fp->f_type != DTYPE_VNODE || vp->v_type == VFIFO) {
+		return (ESPIPE);
+	}
+
+	FREF(fp);
+
+	offset = SCARG(uap, offset);
+
+	/* dofilereadv() will FRELE the descriptor for us */
+	return (dofilereadv(p, fd, fp, SCARG(uap, iovp), SCARG(uap, iovcnt),
+	    &offset, retval));
+}
+
+/*
+ * Positional write system call.
+ */
+int
+sys_pwrite(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_pwrite_args /* {
+		syscallarg(int) fd;
+		syscallarg(const void *) buf;
+		syscallarg(size_t) nbyte;
+		syscallarg(int) pad;
+		syscallarg(off_t) offset;
+	} */ *uap = v;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp;
+	off_t offset;
+	int fd = SCARG(uap, fd);
+
+	if ((fp = fd_getfile(fdp, fd)) == NULL)
+		return (EBADF);
+	if ((fp->f_flag & FWRITE) == 0)
+		return (EBADF);
+
+	vp = (struct vnode *)fp->f_data;
+	if (fp->f_type != DTYPE_VNODE || vp->v_type == VFIFO) {
+		return (ESPIPE);
+	}
+
+	FREF(fp);
+
+	offset = SCARG(uap, offset);
+
+	/* dofilewrite() will FRELE the descriptor for us */
+	return (dofilewrite(p, fd, fp, SCARG(uap, buf), SCARG(uap, nbyte),
+	    &offset, retval));
+}
+
+
+/*
+ * Positional gather write system call.
+ */
+int
+sys_pwritev(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_pwritev_args /* {
+		syscallarg(int) fd;
+		syscallarg(const struct iovec *) iovp;
+		syscallarg(int) iovcnt;
+		syscallarg(int) pad;
+		syscallarg(off_t) offset;
+	} */ *uap = v;
+	struct filedesc *fdp = p->p_fd;
+	struct file *fp;
+	struct vnode *vp;
+	off_t offset;
+	int fd = SCARG(uap, fd);
+
+	if ((fp = fd_getfile(fdp, fd)) == NULL)
+		return (EBADF);
+	if ((fp->f_flag & FWRITE) == 0)
+		return (EBADF);
+
+	vp = (struct vnode *)fp->f_data;
+	if (fp->f_type != DTYPE_VNODE || vp->v_type == VFIFO) {
+		return (ESPIPE);
+	}
+
+	FREF(fp);
+
+	offset = SCARG(uap, offset);
+
+	/* dofilewritev() will FRELE the descriptor for us */
+	return (dofilewritev(p, fd, fp, SCARG(uap, iovp), SCARG(uap, iovcnt),
+	    &offset, retval));
+}
+
+#ifdef UFS_EXTATTR
+/*
+ * 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
+sys_extattrctl(struct proc *p, void *v, register_t *reval)
+{
+	struct sys_extattrctl_args /* {
+		syscallarg(const char *) path;
+		syscallarg(int) cmd;
+		syscallarg(const char *) filename;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+	} */ *uap = v;
+	struct vnode *filename_vp;
+	struct nameidata nd;
+	struct mount *mp;
+	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), p);
+		if ((error = namei(&nd)) != 0)
+			return (error);
+		filename_vp = nd.ni_vp;
+	}
+
+	/* SCARG(uap, path) always defined. */
+	NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, path), p);
+	if ((error = namei(&nd)) != 0) {
+		if (filename_vp != NULL)
+			vput(filename_vp);
+		return (error);
+	}
+
+	mp = nd.ni_vp->v_mount;
+	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, p);
+	} else {
+		error = VFS_EXTATTRCTL(mp, SCARG(uap, cmd), filename_vp,
+		    SCARG(uap, attrnamespace), NULL, p);
+	}
+
+	/*
+	 * 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 proc *p, register_t *retval)
+{
+	struct uio auio;
+	struct iovec aiov;
+	ssize_t cnt;
+	int error;
+
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+
+	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_procp = p;
+	cnt = nbytes;
+
+	error = VOP_SETEXTATTR(vp, attrnamespace, attrname, &auio,
+	    p->p_ucred, p);
+	cnt -= auio.uio_resid;
+	retval[0] = cnt;
+
+done:
+	VOP_UNLOCK(vp, 0, p);
+	return (error);
+}
+
+int
+sys_extattr_set_file(struct proc *p, void *v, register_t *retval)
+{
+	struct sys_extattr_set_file_args /* {
+		syscallarg(const char *) path;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+		syscallarg(void *) data;
+		syscallarg(size_t) nbytes;
+	} */ *uap = v;
+	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), p);
+	if ((error = namei(&nd)) != 0)
+		return (error);
+
+	error = extattr_set_vp(nd.ni_vp, SCARG(uap, attrnamespace), attrname,
+	    SCARG(uap, data), SCARG(uap, nbytes), p, retval);
+
+	vrele(nd.ni_vp);
+	return (error);
+}
+
+int
+sys_extattr_set_fd(struct proc *p, void *v, register_t *retval)
+{
+	struct sys_extattr_set_fd_args /* {
+		syscallarg(int) fd;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+		syscallarg(struct iovec *) iovp;
+		syscallarg(int) iovcnt;
+	} */ *uap = v;
+	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(p->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), p, retval);
+	FRELE(fp);
+
+	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 proc *p, register_t *retval)
+{
+	struct uio auio;
+	struct iovec aiov;
+	ssize_t cnt;
+	size_t size;
+	int error;
+
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_READ);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+
+	/*
+	 * 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_procp = p;
+		cnt = nbytes;
+		error = VOP_GETEXTATTR(vp, attrnamespace, attrname, &auio,
+		    NULL, p->p_ucred, p);
+		cnt -= auio.uio_resid;
+		retval[0] = cnt;
+	} else {
+		error = VOP_GETEXTATTR(vp, attrnamespace, attrname, NULL,
+		    &size, p->p_ucred, p);
+		retval[0] = size;
+	}
+done:
+	VOP_UNLOCK(vp, 0, p);
+	return (error);
+}
+
+int
+sys_extattr_get_file(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_extattr_get_file_args /* {
+		syscallarg(const char *) path;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+		syscallarg(void *) data;
+		syscallarg(size_t) nbytes;
+	} */ *uap = v;
+	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), p);
+	if ((error = namei(&nd)) != 0)
+		return (error);
+
+	error = extattr_get_vp(nd.ni_vp, SCARG(uap, attrnamespace), attrname,
+	    SCARG(uap, data), SCARG(uap, nbytes), p, retval);
+
+	vrele(nd.ni_vp);
+	return (error);
+}
+
+int
+sys_extattr_get_fd(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_extattr_get_fd_args /* {
+		syscallarg(int) fd;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+		syscallarg(void *) data;
+		syscallarg(size_t) nbytes;
+	} */ *uap = v;
+	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(p->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), p, retval);
+	FRELE(fp);
+
+	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 proc *p)
+{
+	int error;
+
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
+	vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
+
+	error = VOP_SETEXTATTR(vp, attrnamespace, attrname, NULL,
+	    p->p_ucred, p);
+
+	VOP_UNLOCK(vp, 0, p);
+	return (error);
+}
+
+int
+sys_extattr_delete_file(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_extattr_delete_file_args /* {
+		syscallarg(int) fd;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+	} */ *uap = v;
+	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), p);
+	if ((error = namei(&nd)) != 0)
+		return(error);
+
+	error = extattr_delete_vp(nd.ni_vp, SCARG(uap, attrnamespace),
+	    attrname, p);
+
+	vrele(nd.ni_vp);
+	return(error);
+}
+
+int
+sys_extattr_delete_fd(p, v, retval)
+	struct proc *p;
+	void *v;
+	register_t *retval;
+{
+	struct sys_extattr_delete_fd_args /* {
+		syscallarg(int) fd;
+		syscallarg(int) attrnamespace;
+		syscallarg(const char *) attrname;
+	} */ *uap = v; 
+	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(p->p_fd, SCARG(uap, fd), &fp)) != 0)
+		return (error);
+
+	error = extattr_delete_vp((struct vnode *)fp->f_data,
+	    SCARG(uap, attrnamespace), attrname, p);
+	FRELE(fp);
+
+	return (error);
+}
+#endif