Merge branch 'work.sane_pwd' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull misc vfs updates from Al Viro:
 "Making sure that something like a referral point won't end up as pwd
  or root.

  The main part is the last commit (fixing mntns_install()); that one
  fixes a hard-to-hit race. The fchdir() commit is making fchdir(2) a
  bit more robust - it should be impossible to get opened files (even
  O_PATH ones) for referral points in the first place, so the existing
  checks are OK, but checking the same thing as in chdir(2) is just as
  cheap.

  The path_init() commit removes a redundant check that shouldn't have
  been there in the first place"

* 'work.sane_pwd' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  make sure that mntns_install() doesn't end up with referral for root
  path_init(): don't bother with checking MAY_EXEC for LOOKUP_ROOT
  make sure that fchdir() won't accept referral points, etc.

fs/namei.c

@@ -2142,7 +2142,6 @@ static int link_path_walk(const char *name, struct nameidata *nd)
 
 static const char *path_init(struct nameidata *nd, unsigned flags)
 {
-	int retval = 0;
 	const char *s = nd->name->name;
 
 	if (!*s)
@@ -2154,13 +2153,8 @@ static const char *path_init(struct nameidata *nd, unsigned flags)
 	if (flags & LOOKUP_ROOT) {
 		struct dentry *root = nd->root.dentry;
 		struct inode *inode = root->d_inode;
-		if (*s) {
-			if (!d_can_lookup(root))
+		if (*s && unlikely(!d_can_lookup(root)))
 			return ERR_PTR(-ENOTDIR);
-			retval = inode_permission(inode, MAY_EXEC);
-			if (retval)
-				return ERR_PTR(retval);
-		}
 		nd->path = nd->root;
 		nd->inode = inode;
 		if (flags & LOOKUP_RCU) {
@@ -2258,6 +2252,35 @@ static inline int lookup_last(struct nameidata *nd)
 	return walk_component(nd, 0);
 }
 
+static int handle_lookup_down(struct nameidata *nd)
+{
+	struct path path = nd->path;
+	struct inode *inode = nd->inode;
+	unsigned seq = nd->seq;
+	int err;
+
+	if (nd->flags & LOOKUP_RCU) {
+		/*
+		 * don't bother with unlazy_walk on failure - we are
+		 * at the very beginning of walk, so we lose nothing
+		 * if we simply redo everything in non-RCU mode
+		 */
+		if (unlikely(!__follow_mount_rcu(nd, &path, &inode, &seq)))
+			return -ECHILD;
+	} else {
+		dget(path.dentry);
+		err = follow_managed(&path, nd);
+		if (unlikely(err < 0))
+			return err;
+		inode = d_backing_inode(path.dentry);
+		seq = 0;
+	}
+	path_to_nameidata(&path, nd);
+	nd->inode = inode;
+	nd->seq = seq;
+	return 0;
+}
+
 /* Returns 0 and nd will be valid on success; Retuns error, otherwise. */
 static int path_lookupat(struct nameidata *nd, unsigned flags, struct path *path)
 {
@@ -2266,6 +2289,15 @@ static int path_lookupat(struct nameidata *nd, unsigned flags, struct path *path
 
 	if (IS_ERR(s))
 		return PTR_ERR(s);
+
+	if (unlikely(flags & LOOKUP_DOWN)) {
+		err = handle_lookup_down(nd);
+		if (unlikely(err < 0)) {
+			terminate_walk(nd);
+			return err;
+		}
+	}
+
 	while (!(err = link_path_walk(s, nd))
 		&& ((err = lookup_last(nd)) > 0)) {
 		s = trailing_symlink(nd);

fs/namespace.c

@@ -3462,8 +3462,9 @@ static void mntns_put(struct ns_common *ns)
 static int mntns_install(struct nsproxy *nsproxy, struct ns_common *ns)
 {
 	struct fs_struct *fs = current->fs;
-	struct mnt_namespace *mnt_ns = to_mnt_ns(ns);
+	struct mnt_namespace *mnt_ns = to_mnt_ns(ns), *old_mnt_ns;
 	struct path root;
+	int err;
 
 	if (!ns_capable(mnt_ns->user_ns, CAP_SYS_ADMIN) ||
 	    !ns_capable(current_user_ns(), CAP_SYS_CHROOT) ||
@@ -3474,15 +3475,18 @@ static int mntns_install(struct nsproxy *nsproxy, struct ns_common *ns)
 		return -EINVAL;
 
 	get_mnt_ns(mnt_ns);
-	put_mnt_ns(nsproxy->mnt_ns);
+	old_mnt_ns = nsproxy->mnt_ns;
 	nsproxy->mnt_ns = mnt_ns;
 
 	/* Find the root */
-	root.mnt    = &mnt_ns->root->mnt;
-	root.dentry = mnt_ns->root->mnt.mnt_root;
-	path_get(&root);
-	while(d_mountpoint(root.dentry) && follow_down_one(&root))
-		;
+	err = vfs_path_lookup(mnt_ns->root->mnt.mnt_root, &mnt_ns->root->mnt,
+				"/", LOOKUP_DOWN, &root);
+	if (err) {
+		/* revert to old namespace */
+		nsproxy->mnt_ns = old_mnt_ns;
+		put_mnt_ns(mnt_ns);
+		return err;
+	}
 
 	/* Update the pwd and root */
 	set_fs_pwd(fs, &root);

fs/open.c

@@ -460,20 +460,17 @@ SYSCALL_DEFINE1(chdir, const char __user *, filename)
 SYSCALL_DEFINE1(fchdir, unsigned int, fd)
 {
 	struct fd f = fdget_raw(fd);
-	struct inode *inode;
-	int error = -EBADF;
+	int error;
 
 	error = -EBADF;
 	if (!f.file)
 		goto out;
 
-	inode = file_inode(f.file);
-
 	error = -ENOTDIR;
-	if (!S_ISDIR(inode->i_mode))
+	if (!d_can_lookup(f.file->f_path.dentry))
 		goto out_putf;
 
-	error = inode_permission(inode, MAY_EXEC | MAY_CHDIR);
+	error = inode_permission(file_inode(f.file), MAY_EXEC | MAY_CHDIR);
 	if (!error)
 		set_fs_pwd(current->fs, &f.file->f_path);
 out_putf:

include/linux/namei.h

@@ -44,6 +44,7 @@ enum {LAST_NORM, LAST_ROOT, LAST_DOT, LAST_DOTDOT, LAST_BIND};
 #define LOOKUP_JUMPED		0x1000
 #define LOOKUP_ROOT		0x2000
 #define LOOKUP_EMPTY		0x4000
+#define LOOKUP_DOWN		0x8000
 
 extern int path_pts(struct path *path);