1 files changed, 196 insertions, 126 deletions

diff --git a/fs/namei.c b/fs/namei.c
index caa28051e197..e029a4cbff7d 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -482,18 +482,6 @@ EXPORT_SYMBOL(path_put);
  * to restart the path walk from the beginning in ref-walk mode.
  */
 
-static inline void lock_rcu_walk(void)
-{
-	br_read_lock(&vfsmount_lock);
-	rcu_read_lock();
-}
-
-static inline void unlock_rcu_walk(void)
-{
-	rcu_read_unlock();
-	br_read_unlock(&vfsmount_lock);
-}
-
 /**
  * unlazy_walk - try to switch to ref-walk mode.
  * @nd: nameidata pathwalk data
@@ -512,26 +500,23 @@ static int unlazy_walk(struct nameidata *nd, struct dentry *dentry)
 	BUG_ON(!(nd->flags & LOOKUP_RCU));
 
 	/*
-	 * Get a reference to the parent first: we're
-	 * going to make "path_put(nd->path)" valid in
-	 * non-RCU context for "terminate_walk()".
-	 *
-	 * If this doesn't work, return immediately with
-	 * RCU walking still active (and then we will do
-	 * the RCU walk cleanup in terminate_walk()).
+	 * After legitimizing the bastards, terminate_walk()
+	 * will do the right thing for non-RCU mode, and all our
+	 * subsequent exit cases should rcu_read_unlock()
+	 * before returning.  Do vfsmount first; if dentry
+	 * can't be legitimized, just set nd->path.dentry to NULL
+	 * and rely on dput(NULL) being a no-op.
 	 */
-	if (!lockref_get_not_dead(&parent->d_lockref))
+	if (!legitimize_mnt(nd->path.mnt, nd->m_seq))
 		return -ECHILD;
-
-	/*
-	 * After the mntget(), we terminate_walk() will do
-	 * the right thing for non-RCU mode, and all our
-	 * subsequent exit cases should unlock_rcu_walk()
-	 * before returning.
-	 */
-	mntget(nd->path.mnt);
 	nd->flags &= ~LOOKUP_RCU;
 
+	if (!lockref_get_not_dead(&parent->d_lockref)) {
+		nd->path.dentry = NULL;	
+		rcu_read_unlock();
+		return -ECHILD;
+	}
+
 	/*
 	 * For a negative lookup, the lookup sequence point is the parents
 	 * sequence point, and it only needs to revalidate the parent dentry.
@@ -566,17 +551,17 @@ static int unlazy_walk(struct nameidata *nd, struct dentry *dentry)
 		spin_unlock(&fs->lock);
 	}
 
-	unlock_rcu_walk();
+	rcu_read_unlock();
 	return 0;
 
 unlock_and_drop_dentry:
 	spin_unlock(&fs->lock);
 drop_dentry:
-	unlock_rcu_walk();
+	rcu_read_unlock();
 	dput(dentry);
 	goto drop_root_mnt;
 out:
-	unlock_rcu_walk();
+	rcu_read_unlock();
 drop_root_mnt:
 	if (!(nd->flags & LOOKUP_ROOT))
 		nd->root.mnt = NULL;
@@ -608,17 +593,22 @@ static int complete_walk(struct nameidata *nd)
 		if (!(nd->flags & LOOKUP_ROOT))
 			nd->root.mnt = NULL;
 
+		if (!legitimize_mnt(nd->path.mnt, nd->m_seq)) {
+			rcu_read_unlock();
+			return -ECHILD;
+		}
 		if (unlikely(!lockref_get_not_dead(&dentry->d_lockref))) {
-			unlock_rcu_walk();
+			rcu_read_unlock();
+			mntput(nd->path.mnt);
 			return -ECHILD;
 		}
 		if (read_seqcount_retry(&dentry->d_seq, nd->seq)) {
-			unlock_rcu_walk();
+			rcu_read_unlock();
 			dput(dentry);
+			mntput(nd->path.mnt);
 			return -ECHILD;
 		}
-		mntget(nd->path.mnt);
-		unlock_rcu_walk();
+		rcu_read_unlock();
 	}
 
 	if (likely(!(nd->flags & LOOKUP_JUMPED)))
@@ -909,15 +899,15 @@ int follow_up(struct path *path)
 	struct mount *parent;
 	struct dentry *mountpoint;
 
-	br_read_lock(&vfsmount_lock);
+	read_seqlock_excl(&mount_lock);
 	parent = mnt->mnt_parent;
 	if (parent == mnt) {
-		br_read_unlock(&vfsmount_lock);
+		read_sequnlock_excl(&mount_lock);
 		return 0;
 	}
 	mntget(&parent->mnt);
 	mountpoint = dget(mnt->mnt_mountpoint);
-	br_read_unlock(&vfsmount_lock);
+	read_sequnlock_excl(&mount_lock);
 	dput(path->dentry);
 	path->dentry = mountpoint;
 	mntput(path->mnt);
@@ -1048,8 +1038,8 @@ static int follow_managed(struct path *path, unsigned flags)
 
 			/* Something is mounted on this dentry in another
 			 * namespace and/or whatever was mounted there in this
-			 * namespace got unmounted before we managed to get the
-			 * vfsmount_lock */
+			 * namespace got unmounted before lookup_mnt() could
+			 * get it */
 		}
 
 		/* Handle an automount point */
@@ -1111,7 +1101,7 @@ static bool __follow_mount_rcu(struct nameidata *nd, struct path *path,
 		if (!d_mountpoint(path->dentry))
 			break;
 
-		mounted = __lookup_mnt(path->mnt, path->dentry, 1);
+		mounted = __lookup_mnt(path->mnt, path->dentry);
 		if (!mounted)
 			break;
 		path->mnt = &mounted->mnt;
@@ -1132,7 +1122,7 @@ static void follow_mount_rcu(struct nameidata *nd)
 {
 	while (d_mountpoint(nd->path.dentry)) {
 		struct mount *mounted;
-		mounted = __lookup_mnt(nd->path.mnt, nd->path.dentry, 1);
+		mounted = __lookup_mnt(nd->path.mnt, nd->path.dentry);
 		if (!mounted)
 			break;
 		nd->path.mnt = &mounted->mnt;
@@ -1174,7 +1164,7 @@ failed:
 	nd->flags &= ~LOOKUP_RCU;
 	if (!(nd->flags & LOOKUP_ROOT))
 		nd->root.mnt = NULL;
-	unlock_rcu_walk();
+	rcu_read_unlock();
 	return -ECHILD;
 }
 
@@ -1308,8 +1298,8 @@ static struct dentry *lookup_dcache(struct qstr *name, struct dentry *dir,
 }
 
 /*
- * Call i_op->lookup on the dentry.  The dentry must be negative but may be
- * hashed if it was pouplated with DCACHE_NEED_LOOKUP.
+ * Call i_op->lookup on the dentry.  The dentry must be negative and
+ * unhashed.
  *
  * dir->d_inode->i_mutex must be held
  */
@@ -1501,7 +1491,7 @@ static void terminate_walk(struct nameidata *nd)
 		nd->flags &= ~LOOKUP_RCU;
 		if (!(nd->flags & LOOKUP_ROOT))
 			nd->root.mnt = NULL;
-		unlock_rcu_walk();
+		rcu_read_unlock();
 	}
 }
 
@@ -1511,18 +1501,9 @@ static void terminate_walk(struct nameidata *nd)
  * so we keep a cache of "no, this doesn't need follow_link"
  * for the common case.
  */
-static inline int should_follow_link(struct inode *inode, int follow)
+static inline int should_follow_link(struct dentry *dentry, int follow)
 {
-	if (unlikely(!(inode->i_opflags & IOP_NOFOLLOW))) {
-		if (likely(inode->i_op->follow_link))
-			return follow;
-
-		/* This gets set once for the inode lifetime */
-		spin_lock(&inode->i_lock);
-		inode->i_opflags |= IOP_NOFOLLOW;
-		spin_unlock(&inode->i_lock);
-	}
-	return 0;
+	return unlikely(d_is_symlink(dentry)) ? follow : 0;
 }
 
 static inline int walk_component(struct nameidata *nd, struct path *path,
@@ -1552,7 +1533,7 @@ static inline int walk_component(struct nameidata *nd, struct path *path,
 	if (!inode)
 		goto out_path_put;
 
-	if (should_follow_link(inode, follow)) {
+	if (should_follow_link(path->dentry, follow)) {
 		if (nd->flags & LOOKUP_RCU) {
 			if (unlikely(unlazy_walk(nd, path->dentry))) {
 				err = -ECHILD;
@@ -1611,26 +1592,6 @@ static inline int nested_symlink(struct path *path, struct nameidata *nd)
 }
 
 /*
- * We really don't want to look at inode->i_op->lookup
- * when we don't have to. So we keep a cache bit in
- * the inode ->i_opflags field that says "yes, we can
- * do lookup on this inode".
- */
-static inline int can_lookup(struct inode *inode)
-{
-	if (likely(inode->i_opflags & IOP_LOOKUP))
-		return 1;
-	if (likely(!inode->i_op->lookup))
-		return 0;
-
-	/* We do this once for the lifetime of the inode */
-	spin_lock(&inode->i_lock);
-	inode->i_opflags |= IOP_LOOKUP;
-	spin_unlock(&inode->i_lock);
-	return 1;
-}
-
-/*
  * We can do the critical dentry name comparison and hashing
  * operations one word at a time, but we are limited to:
  *
@@ -1833,7 +1794,7 @@ static int link_path_walk(const char *name, struct nameidata *nd)
 			if (err)
 				return err;
 		}
-		if (!can_lookup(nd->inode)) {
+		if (!d_is_directory(nd->path.dentry)) {
 			err = -ENOTDIR; 
 			break;
 		}
@@ -1851,9 +1812,10 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 	nd->flags = flags | LOOKUP_JUMPED;
 	nd->depth = 0;
 	if (flags & LOOKUP_ROOT) {
-		struct inode *inode = nd->root.dentry->d_inode;
+		struct dentry *root = nd->root.dentry;
+		struct inode *inode = root->d_inode;
 		if (*name) {
-			if (!can_lookup(inode))
+			if (!d_is_directory(root))
 				return -ENOTDIR;
 			retval = inode_permission(inode, MAY_EXEC);
 			if (retval)
@@ -1862,8 +1824,9 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 		nd->path = nd->root;
 		nd->inode = inode;
 		if (flags & LOOKUP_RCU) {
-			lock_rcu_walk();
+			rcu_read_lock();
 			nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq);
+			nd->m_seq = read_seqbegin(&mount_lock);
 		} else {
 			path_get(&nd->path);
 		}
@@ -1872,9 +1835,10 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 
 	nd->root.mnt = NULL;
 
+	nd->m_seq = read_seqbegin(&mount_lock);
 	if (*name=='/') {
 		if (flags & LOOKUP_RCU) {
-			lock_rcu_walk();
+			rcu_read_lock();
 			set_root_rcu(nd);
 		} else {
 			set_root(nd);
@@ -1886,7 +1850,7 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 			struct fs_struct *fs = current->fs;
 			unsigned seq;
 
-			lock_rcu_walk();
+			rcu_read_lock();
 
 			do {
 				seq = read_seqcount_begin(&fs->seq);
@@ -1907,7 +1871,7 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 		dentry = f.file->f_path.dentry;
 
 		if (*name) {
-			if (!can_lookup(dentry->d_inode)) {
+			if (!d_is_directory(dentry)) {
 				fdput(f);
 				return -ENOTDIR;
 			}
@@ -1918,7 +1882,7 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 			if (f.need_put)
 				*fp = f.file;
 			nd->seq = __read_seqcount_begin(&nd->path.dentry->d_seq);
-			lock_rcu_walk();
+			rcu_read_lock();
 		} else {
 			path_get(&nd->path);
 			fdput(f);
@@ -1989,7 +1953,7 @@ static int path_lookupat(int dfd, const char *name,
 		err = complete_walk(nd);
 
 	if (!err && nd->flags & LOOKUP_DIRECTORY) {
-		if (!can_lookup(nd->inode)) {
+		if (!d_is_directory(nd->path.dentry)) {
 			path_put(&nd->path);
 			err = -ENOTDIR;
 		}
@@ -2281,7 +2245,7 @@ done:
 	}
 	path->dentry = dentry;
 	path->mnt = mntget(nd->path.mnt);
-	if (should_follow_link(dentry->d_inode, nd->flags & LOOKUP_FOLLOW))
+	if (should_follow_link(dentry, nd->flags & LOOKUP_FOLLOW))
 		return 1;
 	follow_mount(path);
 	error = 0;
@@ -2426,12 +2390,14 @@ static inline int check_sticky(struct inode *dir, struct inode *inode)
  * 10. We don't allow removal of NFS sillyrenamed files; it's handled by
  *     nfs_async_unlink().
  */
-static int may_delete(struct inode *dir,struct dentry *victim,int isdir)
+static int may_delete(struct inode *dir, struct dentry *victim, bool isdir)
 {
+	struct inode *inode = victim->d_inode;
 	int error;
 
-	if (!victim->d_inode)
+	if (d_is_negative(victim))
 		return -ENOENT;
+	BUG_ON(!inode);
 
 	BUG_ON(victim->d_parent->d_inode != dir);
 	audit_inode_child(dir, victim, AUDIT_TYPE_CHILD_DELETE);
@@ -2441,15 +2407,16 @@ static int may_delete(struct inode *dir,struct dentry *victim,int isdir)
 		return error;
 	if (IS_APPEND(dir))
 		return -EPERM;
-	if (check_sticky(dir, victim->d_inode)||IS_APPEND(victim->d_inode)||
-	    IS_IMMUTABLE(victim->d_inode) || IS_SWAPFILE(victim->d_inode))
+
+	if (check_sticky(dir, inode) || IS_APPEND(inode) ||
+	    IS_IMMUTABLE(inode) || IS_SWAPFILE(inode))
 		return -EPERM;
 	if (isdir) {
-		if (!S_ISDIR(victim->d_inode->i_mode))
+		if (!d_is_directory(victim) && !d_is_autodir(victim))
 			return -ENOTDIR;
 		if (IS_ROOT(victim))
 			return -EBUSY;
-	} else if (S_ISDIR(victim->d_inode->i_mode))
+	} else if (d_is_directory(victim) || d_is_autodir(victim))
 		return -EISDIR;
 	if (IS_DEADDIR(dir))
 		return -ENOENT;
@@ -2983,7 +2950,7 @@ retry_lookup:
 	/*
 	 * create/update audit record if it already exists.
 	 */
-	if (path->dentry->d_inode)
+	if (d_is_positive(path->dentry))
 		audit_inode(name, path->dentry, 0);
 
 	/*
@@ -3012,12 +2979,12 @@ retry_lookup:
 finish_lookup:
 	/* we _can_ be in RCU mode here */
 	error = -ENOENT;
-	if (!inode) {
+	if (d_is_negative(path->dentry)) {
 		path_to_nameidata(path, nd);
 		goto out;
 	}
 
-	if (should_follow_link(inode, !symlink_ok)) {
+	if (should_follow_link(path->dentry, !symlink_ok)) {
 		if (nd->flags & LOOKUP_RCU) {
 			if (unlikely(unlazy_walk(nd, path->dentry))) {
 				error = -ECHILD;
@@ -3046,10 +3013,11 @@ finish_open:
 	}
 	audit_inode(name, nd->path.dentry, 0);
 	error = -EISDIR;
-	if ((open_flag & O_CREAT) && S_ISDIR(nd->inode->i_mode))
+	if ((open_flag & O_CREAT) &&
+	    (d_is_directory(nd->path.dentry) || d_is_autodir(nd->path.dentry)))
 		goto out;
 	error = -ENOTDIR;
-	if ((nd->flags & LOOKUP_DIRECTORY) && !can_lookup(nd->inode))
+	if ((nd->flags & LOOKUP_DIRECTORY) && !d_is_directory(nd->path.dentry))
 		goto out;
 	if (!S_ISREG(nd->inode->i_mode))
 		will_truncate = false;
@@ -3275,7 +3243,7 @@ struct file *do_file_open_root(struct dentry *dentry, struct vfsmount *mnt,
 	nd.root.mnt = mnt;
 	nd.root.dentry = dentry;
 
-	if (dentry->d_inode->i_op->follow_link && op->intent & LOOKUP_OPEN)
+	if (d_is_symlink(dentry) && op->intent & LOOKUP_OPEN)
 		return ERR_PTR(-ELOOP);
 
 	file = path_openat(-1, &filename, &nd, op, flags | LOOKUP_RCU);
@@ -3325,8 +3293,9 @@ struct dentry *kern_path_create(int dfd, const char *pathname,
 		goto unlock;
 
 	error = -EEXIST;
-	if (dentry->d_inode)
+	if (d_is_positive(dentry))
 		goto fail;
+
 	/*
 	 * Special case - lookup gave negative, but... we had foo/bar/
 	 * From the vfs_mknod() POV we just have a negative dentry -
@@ -3647,8 +3616,27 @@ SYSCALL_DEFINE1(rmdir, const char __user *, pathname)
 	return do_rmdir(AT_FDCWD, pathname);
 }
 
-int vfs_unlink(struct inode *dir, struct dentry *dentry)
+/**
+ * vfs_unlink - unlink a filesystem object
+ * @dir:	parent directory
+ * @dentry:	victim
+ * @delegated_inode: returns victim inode, if the inode is delegated.
+ *
+ * The caller must hold dir->i_mutex.
+ *
+ * If vfs_unlink discovers a delegation, it will return -EWOULDBLOCK and
+ * return a reference to the inode in delegated_inode.  The caller
+ * should then break the delegation on that inode and retry.  Because
+ * breaking a delegation may take a long time, the caller should drop
+ * dir->i_mutex before doing so.
+ *
+ * Alternatively, a caller may pass NULL for delegated_inode.  This may
+ * be appropriate for callers that expect the underlying filesystem not
+ * to be NFS exported.
+ */
+int vfs_unlink(struct inode *dir, struct dentry *dentry, struct inode **delegated_inode)
 {
+	struct inode *target = dentry->d_inode;
 	int error = may_delete(dir, dentry, 0);
 
 	if (error)
@@ -3657,22 +3645,26 @@ int vfs_unlink(struct inode *dir, struct dentry *dentry)
 	if (!dir->i_op->unlink)
 		return -EPERM;
 
-	mutex_lock(&dentry->d_inode->i_mutex);
+	mutex_lock(&target->i_mutex);
 	if (d_mountpoint(dentry))
 		error = -EBUSY;
 	else {
 		error = security_inode_unlink(dir, dentry);
 		if (!error) {
+			error = try_break_deleg(target, delegated_inode);
+			if (error)
+				goto out;
 			error = dir->i_op->unlink(dir, dentry);
 			if (!error)
 				dont_mount(dentry);
 		}
 	}
-	mutex_unlock(&dentry->d_inode->i_mutex);
+out:
+	mutex_unlock(&target->i_mutex);
 
 	/* We don't d_delete() NFS sillyrenamed files--they still exist. */
 	if (!error && !(dentry->d_flags & DCACHE_NFSFS_RENAMED)) {
-		fsnotify_link_count(dentry->d_inode);
+		fsnotify_link_count(target);
 		d_delete(dentry);
 	}
 
@@ -3692,6 +3684,7 @@ static long do_unlinkat(int dfd, const char __user *pathname)
 	struct dentry *dentry;
 	struct nameidata nd;
 	struct inode *inode = NULL;
+	struct inode *delegated_inode = NULL;
 	unsigned int lookup_flags = 0;
 retry:
 	name = user_path_parent(dfd, pathname, &nd, lookup_flags);
@@ -3706,7 +3699,7 @@ retry:
 	error = mnt_want_write(nd.path.mnt);
 	if (error)
 		goto exit1;
-
+retry_deleg:
 	mutex_lock_nested(&nd.path.dentry->d_inode->i_mutex, I_MUTEX_PARENT);
 	dentry = lookup_hash(&nd);
 	error = PTR_ERR(dentry);
@@ -3715,19 +3708,25 @@ retry:
 		if (nd.last.name[nd.last.len])
 			goto slashes;
 		inode = dentry->d_inode;
-		if (!inode)
+		if (d_is_negative(dentry))
 			goto slashes;
 		ihold(inode);
 		error = security_path_unlink(&nd.path, dentry);
 		if (error)
 			goto exit2;
-		error = vfs_unlink(nd.path.dentry->d_inode, dentry);
+		error = vfs_unlink(nd.path.dentry->d_inode, dentry, &delegated_inode);
 exit2:
 		dput(dentry);
 	}
 	mutex_unlock(&nd.path.dentry->d_inode->i_mutex);
 	if (inode)
 		iput(inode);	/* truncate the inode here */
+	inode = NULL;
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
 	mnt_drop_write(nd.path.mnt);
 exit1:
 	path_put(&nd.path);
@@ -3740,8 +3739,12 @@ exit1:
 	return error;
 
 slashes:
-	error = !dentry->d_inode ? -ENOENT :
-		S_ISDIR(dentry->d_inode->i_mode) ? -EISDIR : -ENOTDIR;
+	if (d_is_negative(dentry))
+		error = -ENOENT;
+	else if (d_is_directory(dentry) || d_is_autodir(dentry))
+		error = -EISDIR;
+	else
+		error = -ENOTDIR;
 	goto exit2;
 }
 
@@ -3817,7 +3820,26 @@ SYSCALL_DEFINE2(symlink, const char __user *, oldname, const char __user *, newn
 	return sys_symlinkat(oldname, AT_FDCWD, newname);
 }
 
-int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry)
+/**
+ * vfs_link - create a new link
+ * @old_dentry:	object to be linked
+ * @dir:	new parent
+ * @new_dentry:	where to create the new link
+ * @delegated_inode: returns inode needing a delegation break
+ *
+ * The caller must hold dir->i_mutex
+ *
+ * If vfs_link discovers a delegation on the to-be-linked file in need
+ * of breaking, it will return -EWOULDBLOCK and return a reference to the
+ * inode in delegated_inode.  The caller should then break the delegation
+ * and retry.  Because breaking a delegation may take a long time, the
+ * caller should drop the i_mutex before doing so.
+ *
+ * Alternatively, a caller may pass NULL for delegated_inode.  This may
+ * be appropriate for callers that expect the underlying filesystem not
+ * to be NFS exported.
+ */
+int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_dentry, struct inode **delegated_inode)
 {
 	struct inode *inode = old_dentry->d_inode;
 	unsigned max_links = dir->i_sb->s_max_links;
@@ -3853,8 +3875,11 @@ int vfs_link(struct dentry *old_dentry, struct inode *dir, struct dentry *new_de
 		error =  -ENOENT;
 	else if (max_links && inode->i_nlink >= max_links)
 		error = -EMLINK;
-	else
-		error = dir->i_op->link(old_dentry, dir, new_dentry);
+	else {
+		error = try_break_deleg(inode, delegated_inode);
+		if (!error)
+			error = dir->i_op->link(old_dentry, dir, new_dentry);
+	}
 
 	if (!error && (inode->i_state & I_LINKABLE)) {
 		spin_lock(&inode->i_lock);
@@ -3881,6 +3906,7 @@ SYSCALL_DEFINE5(linkat, int, olddfd, const char __user *, oldname,
 {
 	struct dentry *new_dentry;
 	struct path old_path, new_path;
+	struct inode *delegated_inode = NULL;
 	int how = 0;
 	int error;
 
@@ -3919,9 +3945,14 @@ retry:
 	error = security_path_link(old_path.dentry, &new_path, new_dentry);
 	if (error)
 		goto out_dput;
-	error = vfs_link(old_path.dentry, new_path.dentry->d_inode, new_dentry);
+	error = vfs_link(old_path.dentry, new_path.dentry->d_inode, new_dentry, &delegated_inode);
 out_dput:
 	done_path_create(&new_path, new_dentry);
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry;
+	}
 	if (retry_estale(error, how)) {
 		how |= LOOKUP_REVAL;
 		goto retry;
@@ -3946,7 +3977,8 @@ SYSCALL_DEFINE2(link, const char __user *, oldname, const char __user *, newname
  *	   That's where 4.4 screws up. Current fix: serialization on
  *	   sb->s_vfs_rename_mutex. We might be more accurate, but that's another
  *	   story.
- *	c) we have to lock _three_ objects - parents and victim (if it exists).
+ *	c) we have to lock _four_ objects - parents and victim (if it exists),
+ *	   and source (if it is not a directory).
  *	   And that - after we got ->i_mutex on parents (until then we don't know
  *	   whether the target exists).  Solution: try to be smart with locking
  *	   order for inodes.  We rely on the fact that tree topology may change
@@ -4019,9 +4051,11 @@ out:
 }
 
 static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry,
-			    struct inode *new_dir, struct dentry *new_dentry)
+			    struct inode *new_dir, struct dentry *new_dentry,
+			    struct inode **delegated_inode)
 {
 	struct inode *target = new_dentry->d_inode;
+	struct inode *source = old_dentry->d_inode;
 	int error;
 
 	error = security_inode_rename(old_dir, old_dentry, new_dir, new_dentry);
@@ -4029,13 +4063,20 @@ static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry,
 		return error;
 
 	dget(new_dentry);
-	if (target)
-		mutex_lock(&target->i_mutex);
+	lock_two_nondirectories(source, target);
 
 	error = -EBUSY;
 	if (d_mountpoint(old_dentry)||d_mountpoint(new_dentry))
 		goto out;
 
+	error = try_break_deleg(source, delegated_inode);
+	if (error)
+		goto out;
+	if (target) {
+		error = try_break_deleg(target, delegated_inode);
+		if (error)
+			goto out;
+	}
 	error = old_dir->i_op->rename(old_dir, old_dentry, new_dir, new_dentry);
 	if (error)
 		goto out;
@@ -4045,17 +4086,38 @@ static int vfs_rename_other(struct inode *old_dir, struct dentry *old_dentry,
 	if (!(old_dir->i_sb->s_type->fs_flags & FS_RENAME_DOES_D_MOVE))
 		d_move(old_dentry, new_dentry);
 out:
-	if (target)
-		mutex_unlock(&target->i_mutex);
+	unlock_two_nondirectories(source, target);
 	dput(new_dentry);
 	return error;
 }
 
+/**
+ * vfs_rename - rename a filesystem object
+ * @old_dir:	parent of source
+ * @old_dentry:	source
+ * @new_dir:	parent of destination
+ * @new_dentry:	destination
+ * @delegated_inode: returns an inode needing a delegation break
+ *
+ * The caller must hold multiple mutexes--see lock_rename()).
+ *
+ * If vfs_rename discovers a delegation in need of breaking at either
+ * the source or destination, it will return -EWOULDBLOCK and return a
+ * reference to the inode in delegated_inode.  The caller should then
+ * break the delegation and retry.  Because breaking a delegation may
+ * take a long time, the caller should drop all locks before doing
+ * so.
+ *
+ * Alternatively, a caller may pass NULL for delegated_inode.  This may
+ * be appropriate for callers that expect the underlying filesystem not
+ * to be NFS exported.
+ */
 int vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
-	       struct inode *new_dir, struct dentry *new_dentry)
+	       struct inode *new_dir, struct dentry *new_dentry,
+	       struct inode **delegated_inode)
 {
 	int error;
-	int is_dir = S_ISDIR(old_dentry->d_inode->i_mode);
+	int is_dir = d_is_directory(old_dentry) || d_is_autodir(old_dentry);
 	const unsigned char *old_name;
 
 	if (old_dentry->d_inode == new_dentry->d_inode)
@@ -4080,7 +4142,7 @@ int vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	if (is_dir)
 		error = vfs_rename_dir(old_dir,old_dentry,new_dir,new_dentry);
 	else
-		error = vfs_rename_other(old_dir,old_dentry,new_dir,new_dentry);
+		error = vfs_rename_other(old_dir,old_dentry,new_dir,new_dentry,delegated_inode);
 	if (!error)
 		fsnotify_move(old_dir, new_dir, old_name, is_dir,
 			      new_dentry->d_inode, old_dentry);
@@ -4096,6 +4158,7 @@ SYSCALL_DEFINE4(renameat, int, olddfd, const char __user *, oldname,
 	struct dentry *old_dentry, *new_dentry;
 	struct dentry *trap;
 	struct nameidata oldnd, newnd;
+	struct inode *delegated_inode = NULL;
 	struct filename *from;
 	struct filename *to;
 	unsigned int lookup_flags = 0;
@@ -4135,6 +4198,7 @@ retry:
 	newnd.flags &= ~LOOKUP_PARENT;
 	newnd.flags |= LOOKUP_RENAME_TARGET;
 
+retry_deleg:
 	trap = lock_rename(new_dir, old_dir);
 
 	old_dentry = lookup_hash(&oldnd);
@@ -4143,10 +4207,10 @@ retry:
 		goto exit3;
 	/* source must exist */
 	error = -ENOENT;
-	if (!old_dentry->d_inode)
+	if (d_is_negative(old_dentry))
 		goto exit4;
 	/* unless the source is a directory trailing slashes give -ENOTDIR */
-	if (!S_ISDIR(old_dentry->d_inode->i_mode)) {
+	if (!d_is_directory(old_dentry) && !d_is_autodir(old_dentry)) {
 		error = -ENOTDIR;
 		if (oldnd.last.name[oldnd.last.len])
 			goto exit4;
@@ -4171,13 +4235,19 @@ retry:
 	if (error)
 		goto exit5;
 	error = vfs_rename(old_dir->d_inode, old_dentry,
-				   new_dir->d_inode, new_dentry);
+				   new_dir->d_inode, new_dentry,
+				   &delegated_inode);
 exit5:
 	dput(new_dentry);
 exit4:
 	dput(old_dentry);
 exit3:
 	unlock_rename(new_dir, old_dir);
+	if (delegated_inode) {
+		error = break_deleg_wait(&delegated_inode);
+		if (!error)
+			goto retry_deleg;
+	}
 	mnt_drop_write(oldnd.path.mnt);
 exit2:
 	if (retry_estale(error, lookup_flags))