Merge patch series "Change inode_operations.mkdir to return struct dentry *"

NeilBrown <neilb@suse.de> says:

This revised series contains a few clean-ups as requested by various
people but no substantial changes.

I reviewed the mkdir functions in many (all?) filesystems and found a
few that use d_instantiate() on an unlocked inode (after
unlock_new_inode()) and also support export_operations. These could
potentially call d_instantiate() on a directory inode which is already
attached to a dentry, though making that happen would usually require
guessing the filehandle correctly. I haven't tried to address those
here, (this patch set doesn't make that situation any worse) but I may
in the future.

* patches from https://lore.kernel.org/r/20250227013949.536172-2-neilb@suse.de:
  VFS: Change vfs_mkdir() to return the dentry.
  nfs: change mkdir inode_operation to return alternate dentry if needed.
  fuse: return correct dentry for ->mkdir
  ceph: return the correct dentry on mkdir
  hostfs: store inode in dentry after mkdir if possible.
  Change inode_operations.mkdir to return struct dentry *

Link: https://lore.kernel.org/r/20250227013949.536172-2-neilb@suse.de
Signed-off-by: Christian Brauner <brauner@kernel.org>

Author: brauner

Documentation/filesystems/locking.rst

@@ -66,7 +66,7 @@ prototypes::
 	int (*link) (struct dentry *,struct inode *,struct dentry *);
 	int (*unlink) (struct inode *,struct dentry *);
 	int (*symlink) (struct mnt_idmap *, struct inode *,struct dentry *,const char *);
-	int (*mkdir) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t);
+	struct dentry *(*mkdir) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t);
 	int (*rmdir) (struct inode *,struct dentry *);
 	int (*mknod) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t,dev_t);
 	int (*rename) (struct mnt_idmap *, struct inode *, struct dentry *,

Documentation/filesystems/porting.rst

@@ -1178,3 +1178,22 @@ these conditions don't require explicit checks:
 
 LOOKUP_EXCL now means "target must not exist".  It can be combined with
 LOOK_CREATE or LOOKUP_RENAME_TARGET.
+
+---
+
+** mandatory**
+
+->mkdir() now returns a 'struct dentry *'.  If the created inode is
+found to already be in cache and have a dentry (often IS_ROOT()), it will
+need to be spliced into the given name in place of the given dentry.
+That dentry now needs to be returned.  If the original dentry is used,
+NULL should be returned.  Any error should be returned with
+ERR_PTR().
+
+In general, filesystems which use d_instantiate_new() to install the new
+inode can safely return NULL.  Filesystems which may not have an I_NEW inode
+should use d_drop();d_splice_alias() and return the result of the latter.
+
+If a positive dentry cannot be returned for some reason, in-kernel
+clients such as cachefiles, nfsd, smb/server may not perform ideally but
+will fail-safe.

Documentation/filesystems/vfs.rst

@@ -495,7 +495,7 @@ As of kernel 2.6.22, the following members are defined:
 		int (*link) (struct dentry *,struct inode *,struct dentry *);
 		int (*unlink) (struct inode *,struct dentry *);
 		int (*symlink) (struct mnt_idmap *, struct inode *,struct dentry *,const char *);
-		int (*mkdir) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t);
+		struct dentry *(*mkdir) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t);
 		int (*rmdir) (struct inode *,struct dentry *);
 		int (*mknod) (struct mnt_idmap *, struct inode *,struct dentry *,umode_t,dev_t);
 		int (*rename) (struct mnt_idmap *, struct inode *, struct dentry *,
@@ -562,7 +562,26 @@ otherwise noted.
 ``mkdir``
 	called by the mkdir(2) system call.  Only required if you want
 	to support creating subdirectories.  You will probably need to
-	call d_instantiate() just as you would in the create() method
+	call d_instantiate_new() just as you would in the create() method.
+
+	If d_instantiate_new() is not used and if the fh_to_dentry()
+	export operation is provided, or if the storage might be
+	accessible by another path (e.g. with a network filesystem)
+	then more care may be needed.  Importantly d_instantate()
+	should not be used with an inode that is no longer I_NEW if there
+	any chance that the inode could already be attached to a dentry.
+	This is because of a hard rule in the VFS that a directory must
+	only ever have one dentry.
+
+	For example, if an NFS filesystem is mounted twice the new directory
+	could be visible on the other mount before it is on the original
+	mount, and a pair of name_to_handle_at(), open_by_handle_at()
+	calls could instantiate the directory inode with an IS_ROOT()
+	dentry before the first mkdir returns.
+
+	If there is any chance this could happen, then the new inode
+	should be d_drop()ed and attached with d_splice_alias().  The
+	returned dentry (if any) should be returned by ->mkdir().
 
 ``rmdir``
 	called by the rmdir(2) system call.  Only required if you want

drivers/base/devtmpfs.c

@@ -160,18 +160,17 @@ static int dev_mkdir(const char *name, umode_t mode)
 {
 	struct dentry *dentry;
 	struct path path;
-	int err;
 
 	dentry = kern_path_create(AT_FDCWD, name, &path, LOOKUP_DIRECTORY);
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
-	err = vfs_mkdir(&nop_mnt_idmap, d_inode(path.dentry), dentry, mode);
-	if (!err)
+	dentry = vfs_mkdir(&nop_mnt_idmap, d_inode(path.dentry), dentry, mode);
+	if (!IS_ERR(dentry))
 		/* mark as kernel-created inode */
 		d_inode(dentry)->i_private = &thread;
 	done_path_create(&path, dentry);
-	return err;
+	return PTR_ERR_OR_ZERO(dentry);
 }
 
 static int create_path(const char *nodepath)

fs/9p/vfs_inode.c

@@ -669,8 +669,8 @@ v9fs_vfs_create(struct mnt_idmap *idmap, struct inode *dir,
  *
  */
 
-static int v9fs_vfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
-			  struct dentry *dentry, umode_t mode)
+static struct dentry *v9fs_vfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+				     struct dentry *dentry, umode_t mode)
 {
 	int err;
 	u32 perm;
@@ -692,8 +692,7 @@ static int v9fs_vfs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 
 	if (fid)
 		p9_fid_put(fid);
-
-	return err;
+	return ERR_PTR(err);
 }
 
 /**

fs/9p/vfs_inode_dotl.c

@@ -350,9 +350,9 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry,
  *
  */
 
-static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
-			       struct inode *dir, struct dentry *dentry,
-			       umode_t omode)
+static struct dentry *v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
+					  struct inode *dir, struct dentry *dentry,
+					  umode_t omode)
 {
 	int err;
 	struct v9fs_session_info *v9ses;
@@ -417,7 +417,7 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap,
 	p9_fid_put(fid);
 	v9fs_put_acl(dacl, pacl);
 	p9_fid_put(dfid);
-	return err;
+	return ERR_PTR(err);
 }
 
 static int

fs/affs/affs.h

@@ -168,7 +168,7 @@ extern struct dentry *affs_lookup(struct inode *dir, struct dentry *dentry, unsi
 extern int	affs_unlink(struct inode *dir, struct dentry *dentry);
 extern int	affs_create(struct mnt_idmap *idmap, struct inode *dir,
 			struct dentry *dentry, umode_t mode, bool);
-extern int	affs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+extern struct dentry *affs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 			struct dentry *dentry, umode_t mode);
 extern int	affs_rmdir(struct inode *dir, struct dentry *dentry);
 extern int	affs_link(struct dentry *olddentry, struct inode *dir,

fs/affs/namei.c

@@ -273,7 +273,7 @@ affs_create(struct mnt_idmap *idmap, struct inode *dir,
 	return 0;
 }
 
-int
+struct dentry *
 affs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 	   struct dentry *dentry, umode_t mode)
 {
@@ -285,7 +285,7 @@ affs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 
 	inode = affs_new_inode(dir);
 	if (!inode)
-		return -ENOSPC;
+		return ERR_PTR(-ENOSPC);
 
 	inode->i_mode = S_IFDIR | mode;
 	affs_mode_to_prot(inode);
@@ -298,9 +298,9 @@ affs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 		clear_nlink(inode);
 		mark_inode_dirty(inode);
 		iput(inode);
-		return error;
+		return ERR_PTR(error);
 	}
-	return 0;
+	return NULL;
 }
 
 int

fs/afs/dir.c

@@ -33,8 +33,8 @@ static bool afs_lookup_filldir(struct dir_context *ctx, const char *name, int nl
 			      loff_t fpos, u64 ino, unsigned dtype);
 static int afs_create(struct mnt_idmap *idmap, struct inode *dir,
 		      struct dentry *dentry, umode_t mode, bool excl);
-static int afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
-		     struct dentry *dentry, umode_t mode);
+static struct dentry *afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+				struct dentry *dentry, umode_t mode);
 static int afs_rmdir(struct inode *dir, struct dentry *dentry);
 static int afs_unlink(struct inode *dir, struct dentry *dentry);
 static int afs_link(struct dentry *from, struct inode *dir,
@@ -1315,8 +1315,8 @@ static const struct afs_operation_ops afs_mkdir_operation = {
 /*
  * create a directory on an AFS filesystem
  */
-static int afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
-		     struct dentry *dentry, umode_t mode)
+static struct dentry *afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+				struct dentry *dentry, umode_t mode)
 {
 	struct afs_operation *op;
 	struct afs_vnode *dvnode = AFS_FS_I(dir);
@@ -1328,7 +1328,7 @@ static int afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 	op = afs_alloc_operation(NULL, dvnode->volume);
 	if (IS_ERR(op)) {
 		d_drop(dentry);
-		return PTR_ERR(op);
+		return ERR_CAST(op);
 	}
 
 	fscache_use_cookie(afs_vnode_cache(dvnode), true);
@@ -1344,7 +1344,7 @@ static int afs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
 	op->ops		= &afs_mkdir_operation;
 	ret = afs_do_sync_operation(op);
 	afs_dir_unuse_cookie(dvnode, ret);
-	return ret;
+	return ERR_PTR(ret);
 }
 
 /*

fs/autofs/root.c

@@ -15,8 +15,8 @@ static int autofs_dir_symlink(struct mnt_idmap *, struct inode *,
 			      struct dentry *, const char *);
 static int autofs_dir_unlink(struct inode *, struct dentry *);
 static int autofs_dir_rmdir(struct inode *, struct dentry *);
-static int autofs_dir_mkdir(struct mnt_idmap *, struct inode *,
-			    struct dentry *, umode_t);
+static struct dentry *autofs_dir_mkdir(struct mnt_idmap *, struct inode *,
+				       struct dentry *, umode_t);
 static long autofs_root_ioctl(struct file *, unsigned int, unsigned long);
 #ifdef CONFIG_COMPAT
 static long autofs_root_compat_ioctl(struct file *,
@@ -720,9 +720,9 @@ static int autofs_dir_rmdir(struct inode *dir, struct dentry *dentry)
 	return 0;
 }
 
-static int autofs_dir_mkdir(struct mnt_idmap *idmap,
-			    struct inode *dir, struct dentry *dentry,
-			    umode_t mode)
+static struct dentry *autofs_dir_mkdir(struct mnt_idmap *idmap,
+				       struct inode *dir, struct dentry *dentry,
+				       umode_t mode)
 {
 	struct autofs_sb_info *sbi = autofs_sbi(dir->i_sb);
 	struct autofs_info *ino = autofs_dentry_ino(dentry);
@@ -739,7 +739,7 @@ static int autofs_dir_mkdir(struct mnt_idmap *idmap,
 
 	inode = autofs_get_inode(dir->i_sb, S_IFDIR | mode);
 	if (!inode)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 	d_add(dentry, inode);
 
 	if (sbi->version < 5)
@@ -751,7 +751,7 @@ static int autofs_dir_mkdir(struct mnt_idmap *idmap,
 	inc_nlink(dir);
 	inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir));
 
-	return 0;
+	return NULL;
 }
 
 /* Get/set timeout ioctl() operation */