X-Git-Url: https://git.subgeniuskitty.com/unix-history/.git/blobdiff_plain/b5e5c7928c3c69241fc8e32f6475880c824578d6..1a7bccf6251d2e3a04b364630a80a7f1760cb09b:/usr/src/sys/kern/vfs_syscalls.c

diff --git a/usr/src/sys/kern/vfs_syscalls.c b/usr/src/sys/kern/vfs_syscalls.c
index 3c76b93c28..deb34b982c 100644
--- a/usr/src/sys/kern/vfs_syscalls.c
+++ b/usr/src/sys/kern/vfs_syscalls.c
@@ -9,7 +9,7 @@
  *
  * %sccs.include.redist.c%
  *
- *	@(#)vfs_syscalls.c	8.16 (Berkeley) %G%
+ *	@(#)vfs_syscalls.c	8.21 (Berkeley) %G%
  */
 
 #include <sys/param.h>
@@ -63,13 +63,9 @@ mount(p, uap, retval)
 	register struct vnode *vp;
 	register struct mount *mp;
 	int error, flag;
+	struct vattr va;
 	struct nameidata nd;
 
-	/*
-	 * Must be super user
-	 */
-	if (error = suser(p->p_ucred, &p->p_acflag))
-		return (error);
 	/*
 	 * Get vnode to be covered
 	 */
@@ -95,9 +91,50 @@ mount(p, uap, retval)
 		}
 		mp->mnt_flag |=
 		    uap->flags & (MNT_RELOAD | MNT_FORCE | MNT_UPDATE);
+		/*
+		 * Only root, or the user that did the original mount is
+		 * permitted to update it.
+		 */
+		if (mp->mnt_stat.f_owner != p->p_ucred->cr_uid &&
+		    (error = suser(p->p_ucred, &p->p_acflag))) {
+			vput(vp);
+			return (error);
+		}
+		/*
+		 * Do not allow NFS export by non-root users. Silently
+		 * enforce MNT_NOSUID and MNT_NODEV for non-root users.
+		 */
+		if (p->p_ucred->cr_uid != 0) {
+			if (uap->flags & MNT_EXPORTED) {
+				vput(vp);
+				return (EPERM);
+			}
+			uap->flags |= MNT_NOSUID | MNT_NODEV;
+		}
 		VOP_UNLOCK(vp);
 		goto update;
 	}
+	/*
+	 * If the user is not root, ensure that they own the directory
+	 * onto which we are attempting to mount.
+	 */
+	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)))) {
+		vput(vp);
+		return (error);
+	}
+	/*
+	 * Do not allow NFS export by non-root users. Silently
+	 * enforce MNT_NOSUID and MNT_NODEV for non-root users.
+	 */
+	if (p->p_ucred->cr_uid != 0) {
+		if (uap->flags & MNT_EXPORTED) {
+			vput(vp);
+			return (EPERM);
+		}
+		uap->flags |= MNT_NOSUID | MNT_NODEV;
+	}
 	if (error = vinvalbuf(vp, V_SAVE, p->p_ucred, p, 0, 0))
 		return (error);
 	if (vp->v_type != VDIR) {
@@ -129,6 +166,7 @@ mount(p, uap, retval)
 	}
 	vp->v_mountedhere = mp;
 	mp->mnt_vnodecovered = vp;
+	mp->mnt_stat.f_owner = p->p_ucred->cr_uid;
 update:
 	/*
 	 * Set the mount level flags.
@@ -198,12 +236,13 @@ unmount(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
+	mp = vp->v_mount;
 
 	/*
-	 * Unless this is a user mount, then must
-	 * have suser privilege.
+	 * Only root, or the user that did the original mount is
+	 * permitted to unmount this filesystem.
 	 */
-	if (((vp->v_mount->mnt_flag & MNT_USER) == 0) &&
+	if ((mp->mnt_stat.f_owner != p->p_ucred->cr_uid) &&
 	    (error = suser(p->p_ucred, &p->p_acflag))) {
 		vput(vp);
 		return (error);
@@ -216,7 +255,6 @@ unmount(p, uap, retval)
 		vput(vp);
 		return (EINVAL);
 	}
-	mp = vp->v_mount;
 	vput(vp);
 	return (dounmount(mp, uap->flags, p));
 }
@@ -282,6 +320,10 @@ sync(p, uap, retval)
 	int asyncflag;
 
 	for (mp = mountlist.tqh_first; mp != NULL; mp = nmp) {
+		/*
+		 * Get the next pointer in case we hang on vfs_busy
+		 * while we are being unmounted.
+		 */
 		nmp = mp->mnt_list.tqe_next;
 		/*
 		 * The lock check below is to avoid races with mount
@@ -294,6 +336,11 @@ sync(p, uap, retval)
 			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.tqe_next;
 			vfs_unbusy(mp);
 		}
 	}
@@ -661,6 +708,7 @@ mknod(p, uap, retval)
 	register struct vnode *vp;
 	struct vattr vattr;
 	int error;
+	int whiteout;
 	struct nameidata nd;
 
 	CHECKPOINTREF;
@@ -676,6 +724,7 @@ mknod(p, uap, retval)
 		VATTR_NULL(&vattr);
 		vattr.va_mode = (uap->mode & ALLPERMS) &~ p->p_fd->fd_cmask;
 		vattr.va_rdev = uap->dev;
+		whiteout = 0;
 
 		switch (uap->mode & S_IFMT) {
 		case S_IFMT:	/* used by badsect to flag bad sectors */
@@ -687,13 +736,19 @@ mknod(p, uap, retval)
 		case S_IFBLK:
 			vattr.va_type = VBLK;
 			break;
+		case S_IFWHT:
+			whiteout = 1;
+			break;
 		default:
 			error = EINVAL;
 			break;
 		}
 	}
-	if (!error) {
-		LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	if (whiteout) {
+		error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, CREATE);
+		vput(nd.ni_dvp);
+	} else if (!error) {
+		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
 		error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr);
 	} else {
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
@@ -743,7 +798,7 @@ mkfifo(p, uap, retval)
 	VATTR_NULL(&vattr);
 	vattr.va_type = VFIFO;
 	vattr.va_mode = (uap->mode & ALLPERMS) &~ p->p_fd->fd_cmask;
-	LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	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 */
 }
@@ -779,10 +834,9 @@ link(p, uap, retval)
 			if (nd.ni_vp != NULL)
 				error = EEXIST;
 			if (!error) {
-				LEASE_CHECK(nd.ni_dvp,
-				    p, p->p_ucred, LEASE_WRITE);
-				LEASE_CHECK(vp,
-				    p, p->p_ucred, LEASE_WRITE);
+				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);
@@ -837,7 +891,7 @@ symlink(p, uap, retval)
 	}
 	VATTR_NULL(&vattr);
 	vattr.va_mode = ACCESSPERMS &~ p->p_fd->fd_cmask;
-	LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
 	error = VOP_SYMLINK(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr, path);
 out:
 	FREE(path, M_NAMEI);
@@ -845,6 +899,45 @@ out:
 	return (error);
 }
 
+/*
+ * Delete a whiteout from the filesystem.
+ */
+struct unwhiteout_args {
+	char	*path;
+};
+/* ARGSUSED */
+unwhiteout(p, uap, retval)
+	struct proc *p;
+	struct unwhiteout_args *uap;
+	int *retval;
+{
+	int error;
+	struct nameidata nd;
+
+	NDINIT(&nd, DELETE, LOCKPARENT|DOWHITEOUT, UIO_USERSPACE, uap->path, p);
+	error = namei(&nd);
+	if (error)
+		return (error);
+
+	if (nd.ni_vp != NULLVP || !(nd.ni_cnd.cn_flags & ISWHITEOUT)) {
+		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);
+		return (EEXIST);
+	}
+
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, DELETE);
+	if (error != 0)
+		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
+	vput(nd.ni_dvp);
+	return (error);
+}
+
 /*
  * Delete a name from the filesystem.
  */
@@ -866,7 +959,7 @@ unlink(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 
 	if (vp->v_type != VDIR ||
@@ -881,7 +974,7 @@ unlink(p, uap, retval)
 	}
 
 	if (!error) {
-		LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
 		error = VOP_REMOVE(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd);
 	} else {
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
@@ -889,7 +982,8 @@ unlink(p, uap, retval)
 			vrele(nd.ni_dvp);
 		else
 			vput(nd.ni_dvp);
-		vput(vp);
+		if (vp != NULLVP)
+			vput(vp);
 	}
 	CHECKREFS("unlink");
 	return (error);
@@ -1279,7 +1373,7 @@ chflags(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1313,7 +1407,7 @@ fchflags(p, uap, retval)
 	if (error = getvnode(p->p_fd, uap->fd, &fp))
 		return (error);
 	vp = (struct vnode *)fp->f_data;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1348,7 +1442,7 @@ chmod(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1382,7 +1476,7 @@ fchmod(p, uap, retval)
 	if (error = getvnode(p->p_fd, uap->fd, &fp))
 		return (error);
 	vp = (struct vnode *)fp->f_data;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1418,7 +1512,7 @@ chown(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1454,7 +1548,7 @@ fchown(p, uap, retval)
 	if (error = getvnode(p->p_fd, uap->fd, &fp))
 		return (error);
 	vp = (struct vnode *)fp->f_data;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1498,7 +1592,7 @@ utimes(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_mount->mnt_flag & MNT_RDONLY)
 		error = EROFS;
@@ -1536,7 +1630,7 @@ truncate(p, uap, retval)
 	if (error = namei(&nd))
 		return (error);
 	vp = nd.ni_vp;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_type == VDIR)
 		error = EISDIR;
@@ -1574,7 +1668,7 @@ ftruncate(p, uap, retval)
 	if ((fp->f_flag & FWRITE) == 0)
 		return (EINVAL);
 	vp = (struct vnode *)fp->f_data;
-	LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 	VOP_LOCK(vp);
 	if (vp->v_type == VDIR)
 		error = EISDIR;
@@ -1711,11 +1805,11 @@ rename(p, uap, retval)
 		error = -1;
 out:
 	if (!error) {
-		LEASE_CHECK(tdvp, p, p->p_ucred, LEASE_WRITE);
+		VOP_LEASE(tdvp, p, p->p_ucred, LEASE_WRITE);
 		if (fromnd.ni_dvp != tdvp)
-			LEASE_CHECK(fromnd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+			VOP_LEASE(fromnd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
 		if (tvp)
-			LEASE_CHECK(tvp, p, p->p_ucred, LEASE_WRITE);
+			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 {
@@ -1780,7 +1874,7 @@ mkdir(p, uap, retval)
 	VATTR_NULL(&vattr);
 	vattr.va_type = VDIR;
 	vattr.va_mode = (uap->mode & ACCESSPERMS) &~ p->p_fd->fd_cmask;
-	LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+	VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
 	error = VOP_MKDIR(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr);
 	if (!error)
 		vput(nd.ni_vp);
@@ -1827,8 +1921,8 @@ rmdir(p, uap, retval)
 		error = EBUSY;
 out:
 	if (!error) {
-		LEASE_CHECK(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
-		LEASE_CHECK(vp, p, p->p_ucred, LEASE_WRITE);
+		VOP_LEASE(nd.ni_dvp, p, p->p_ucred, LEASE_WRITE);
+		VOP_LEASE(vp, p, p->p_ucred, LEASE_WRITE);
 		error = VOP_RMDIR(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd);
 	} else {
 		VOP_ABORTOP(nd.ni_dvp, &nd.ni_cnd);
@@ -1949,6 +2043,20 @@ unionread:
 		struct vnode *lvp;
 
 		lvp = union_lowervp(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) {
+				vrele(lvp);
+				lvp = NULL;
+			}
+		}
+		
 		if (lvp != NULLVP) {
 			VOP_LOCK(lvp);
 			error = VOP_OPEN(lvp, FREAD, fp->f_cred, p);
@@ -2042,6 +2150,20 @@ unionread:
 		struct vnode *lvp;
 
 		lvp = union_lowervp(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) {
+				vrele(lvp);
+				lvp = NULL;
+			}
+		}
+		
 		if (lvp != NULLVP) {
 			VOP_LOCK(lvp);
 			error = VOP_OPEN(lvp, FREAD, fp->f_cred, p);