Merge tag 'ovl-update-6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs

Pull overlayfs update from Miklos Szeredi:

 - Fix a couple of bugs found by syzbot

 - Don't ingore some open flags set by fcntl(F_SETFL)

 - Fix failure to create a hard link in certain cases

 - Use type safe helpers for some mnt_userns transformations

 - Improve performance of mount

 - Misc cleanups

* tag 'ovl-update-6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs:
  ovl: Kconfig: Fix spelling mistake "undelying" -> "underlying"
  ovl: use inode instead of dentry where possible
  ovl: Add comment on upperredirect reassignment
  ovl: use plain list filler in indexdir and workdir cleanup
  ovl: do not reconnect upper index records in ovl_indexdir_cleanup()
  ovl: fix comment typos
  ovl: port to vfs{g,u}id_t and associated helpers
  ovl: Use ovl mounter's fsuid and fsgid in ovl_link()
  ovl: Use "buf" flexible array for memcpy() destination
  ovl: update ->f_iocb_flags when ovl_change_flags() modifies ->f_flags
  ovl: fix use inode directly in rcu-walk mode

9 files changed, 86 insertions, 67 deletions

diff --git a/fs/overlayfs/Kconfig b/fs/overlayfs/Kconfig
index dd188c7996b318..6708e54b0e30a4 100644
--- a/fs/overlayfs/Kconfig
+++ b/fs/overlayfs/Kconfig
@@ -96,7 +96,7 @@ config OVERLAY_FS_XINO_AUTO
 	depends on 64BIT
 	help
 	  If this config option is enabled then overlay filesystems will use
-	  unused high bits in undelying filesystem inode numbers to map all
+	  unused high bits in underlying filesystem inode numbers to map all
 	  inodes to a unified address space.  The mapped 64bit inode numbers
 	  might not be compatible with applications that expect 32bit inodes.
 
diff --git a/fs/overlayfs/dir.c b/fs/overlayfs/dir.c
index cbb569d5d234ae..f61e37f4c8ff4c 100644
--- a/fs/overlayfs/dir.c
+++ b/fs/overlayfs/dir.c
@@ -576,28 +576,42 @@ static int ovl_create_or_link(struct dentry *dentry, struct inode *inode,
 			goto out_revert_creds;
 	}
 
-	err = -ENOMEM;
-	override_cred = prepare_creds();
-	if (override_cred) {
+	if (!attr->hardlink) {
+		err = -ENOMEM;
+		override_cred = prepare_creds();
+		if (!override_cred)
+			goto out_revert_creds;
+		/*
+		 * In the creation cases(create, mkdir, mknod, symlink),
+		 * ovl should transfer current's fs{u,g}id to underlying
+		 * fs. Because underlying fs want to initialize its new
+		 * inode owner using current's fs{u,g}id. And in this
+		 * case, the @inode is a new inode that is initialized
+		 * in inode_init_owner() to current's fs{u,g}id. So use
+		 * the inode's i_{u,g}id to override the cred's fs{u,g}id.
+		 *
+		 * But in the other hardlink case, ovl_link() does not
+		 * create a new inode, so just use the ovl mounter's
+		 * fs{u,g}id.
+		 */
 		override_cred->fsuid = inode->i_uid;
 		override_cred->fsgid = inode->i_gid;
-		if (!attr->hardlink) {
-			err = security_dentry_create_files_as(dentry,
-					attr->mode, &dentry->d_name, old_cred,
-					override_cred);
-			if (err) {
-				put_cred(override_cred);
-				goto out_revert_creds;
-			}
+		err = security_dentry_create_files_as(dentry,
+				attr->mode, &dentry->d_name, old_cred,
+				override_cred);
+		if (err) {
+			put_cred(override_cred);
+			goto out_revert_creds;
 		}
 		put_cred(override_creds(override_cred));
 		put_cred(override_cred);
-
-		if (!ovl_dentry_is_whiteout(dentry))
-			err = ovl_create_upper(dentry, inode, attr);
-		else
-			err = ovl_create_over_whiteout(dentry, inode, attr);
 	}
+
+	if (!ovl_dentry_is_whiteout(dentry))
+		err = ovl_create_upper(dentry, inode, attr);
+	else
+		err = ovl_create_over_whiteout(dentry, inode, attr);
+
 out_revert_creds:
 	revert_creds(old_cred);
 	return err;
diff --git a/fs/overlayfs/export.c b/fs/overlayfs/export.c
index e065a5b9a442e2..a25bb3453ddeca 100644
--- a/fs/overlayfs/export.c
+++ b/fs/overlayfs/export.c
@@ -339,7 +339,7 @@ out_iput:
 	return dentry;
 }
 
-/* Get the upper or lower dentry in stach whose on layer @idx */
+/* Get the upper or lower dentry in stack whose on layer @idx */
 static struct dentry *ovl_dentry_real_at(struct dentry *dentry, int idx)
 {
 	struct ovl_entry *oe = dentry->d_fsdata;
@@ -463,7 +463,7 @@ static struct dentry *ovl_lookup_real_inode(struct super_block *sb,
 
 	/* Get connected upper overlay dir from index */
 	if (index) {
-		struct dentry *upper = ovl_index_upper(ofs, index);
+		struct dentry *upper = ovl_index_upper(ofs, index, true);
 
 		dput(index);
 		if (IS_ERR_OR_NULL(upper))
@@ -739,7 +739,7 @@ static struct dentry *ovl_lower_fh_to_d(struct super_block *sb,
 
 	/* Then try to get a connected upper dir by index */
 	if (index && d_is_dir(index)) {
-		struct dentry *upper = ovl_index_upper(ofs, index);
+		struct dentry *upper = ovl_index_upper(ofs, index, true);
 
 		err = PTR_ERR(upper);
 		if (IS_ERR_OR_NULL(upper))
@@ -796,7 +796,7 @@ static struct ovl_fh *ovl_fid_to_fh(struct fid *fid, int buflen, int fh_type)
 		return ERR_PTR(-ENOMEM);
 
 	/* Copy unaligned inner fh into aligned buffer */
-	memcpy(&fh->fb, fid, buflen - OVL_FH_WIRE_OFFSET);
+	memcpy(fh->buf, fid, buflen - OVL_FH_WIRE_OFFSET);
 	return fh;
 }
 
diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
index d066be3b9226ec..c9d0c362c7ef11 100644
--- a/fs/overlayfs/file.c
+++ b/fs/overlayfs/file.c
@@ -34,7 +34,7 @@ static char ovl_whatisit(struct inode *inode, struct inode *realinode)
 		return 'm';
 }
 
-/* No atime modificaton nor notify on underlying */
+/* No atime modification nor notify on underlying */
 #define OVL_OPEN_FLAGS (O_NOATIME | FMODE_NONOTIFY)
 
 static struct file *ovl_open_realfile(const struct file *file,
@@ -96,6 +96,7 @@ static int ovl_change_flags(struct file *file, unsigned int flags)
 
 	spin_lock(&file->f_lock);
 	file->f_flags = (file->f_flags & ~OVL_SETFL_MASK) | flags;
+	file->f_iocb_flags = iocb_flags(file);
 	spin_unlock(&file->f_lock);
 
 	return 0;
diff --git a/fs/overlayfs/namei.c b/fs/overlayfs/namei.c
index 0fd1d5fdfc728b..46753134533a1e 100644
--- a/fs/overlayfs/namei.c
+++ b/fs/overlayfs/namei.c
@@ -487,7 +487,8 @@ fail:
 }
 
 /* Get upper dentry from index */
-struct dentry *ovl_index_upper(struct ovl_fs *ofs, struct dentry *index)
+struct dentry *ovl_index_upper(struct ovl_fs *ofs, struct dentry *index,
+			       bool connected)
 {
 	struct ovl_fh *fh;
 	struct dentry *upper;
@@ -499,7 +500,7 @@ struct dentry *ovl_index_upper(struct ovl_fs *ofs, struct dentry *index)
 	if (IS_ERR_OR_NULL(fh))
 		return ERR_CAST(fh);
 
-	upper = ovl_decode_real_fh(ofs, fh, ovl_upper_mnt(ofs), true);
+	upper = ovl_decode_real_fh(ofs, fh, ovl_upper_mnt(ofs), connected);
 	kfree(fh);
 
 	if (IS_ERR_OR_NULL(upper))
@@ -572,7 +573,7 @@ int ovl_verify_index(struct ovl_fs *ofs, struct dentry *index)
 	 * directly from the index dentry, but for dir index we first need to
 	 * decode the upper directory.
 	 */
-	upper = ovl_index_upper(ofs, index);
+	upper = ovl_index_upper(ofs, index, false);
 	if (IS_ERR_OR_NULL(upper)) {
 		err = PTR_ERR(upper);
 		/*
@@ -1085,6 +1086,11 @@ struct dentry *ovl_lookup(struct inode *dir, struct dentry *dentry,
 			.mnt = ovl_upper_mnt(ofs),
 		};
 
+		/*
+		 * It's safe to assign upperredirect here: the previous
+		 * assignment of happens only if upperdentry is non-NULL, and
+		 * this one only if upperdentry is NULL.
+		 */
 		upperredirect = ovl_get_redirect_xattr(ofs, &upperpath, 0);
 		if (IS_ERR(upperredirect)) {
 			err = PTR_ERR(upperredirect);
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index 480e6aabef27a8..1df7f850ff3bb2 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -110,7 +110,7 @@ struct ovl_fh {
 	u8 padding[3];	/* make sure fb.fid is 32bit aligned */
 	union {
 		struct ovl_fb fb;
-		u8 buf[0];
+		DECLARE_FLEX_ARRAY(u8, buf);
 	};
 } __packed;
 
@@ -415,7 +415,7 @@ const char *ovl_dentry_get_redirect(struct dentry *dentry);
 void ovl_dentry_set_redirect(struct dentry *dentry, const char *redirect);
 void ovl_inode_update(struct inode *inode, struct dentry *upperdentry);
 void ovl_dir_modified(struct dentry *dentry, bool impurity);
-u64 ovl_dentry_version_get(struct dentry *dentry);
+u64 ovl_inode_version_get(struct inode *inode);
 bool ovl_is_whiteout(struct dentry *dentry);
 struct file *ovl_path_open(const struct path *path, int flags);
 int ovl_copy_up_start(struct dentry *dentry, int flags);
@@ -539,7 +539,8 @@ int ovl_check_origin_fh(struct ovl_fs *ofs, struct ovl_fh *fh, bool connected,
 int ovl_verify_set_fh(struct ovl_fs *ofs, struct dentry *dentry,
 		      enum ovl_xattr ox, struct dentry *real, bool is_upper,
 		      bool set);
-struct dentry *ovl_index_upper(struct ovl_fs *ofs, struct dentry *index);
+struct dentry *ovl_index_upper(struct ovl_fs *ofs, struct dentry *index,
+			       bool connected);
 int ovl_verify_index(struct ovl_fs *ofs, struct dentry *index);
 int ovl_get_index_name(struct ovl_fs *ofs, struct dentry *origin,
 		       struct qstr *name);
@@ -584,9 +585,9 @@ int ovl_indexdir_cleanup(struct ovl_fs *ofs);
  * lower dir was removed under it and possibly before it was rotated from upper
  * to lower layer.
  */
-static inline bool ovl_dir_is_real(struct dentry *dir)
+static inline bool ovl_dir_is_real(struct inode *dir)
 {
-	return !ovl_test_flag(OVL_WHITEOUTS, d_inode(dir));
+	return !ovl_test_flag(OVL_WHITEOUTS, dir);
 }
 
 /* inode.c */
diff --git a/fs/overlayfs/readdir.c b/fs/overlayfs/readdir.c
index 2b210640036c4b..8cd2b9947de1de 100644
--- a/fs/overlayfs/readdir.c
+++ b/fs/overlayfs/readdir.c
@@ -235,15 +235,15 @@ void ovl_dir_cache_free(struct inode *inode)
 	}
 }
 
-static void ovl_cache_put(struct ovl_dir_file *od, struct dentry *dentry)
+static void ovl_cache_put(struct ovl_dir_file *od, struct inode *inode)
 {
 	struct ovl_dir_cache *cache = od->cache;
 
 	WARN_ON(cache->refcount <= 0);
 	cache->refcount--;
 	if (!cache->refcount) {
-		if (ovl_dir_cache(d_inode(dentry)) == cache)
-			ovl_set_dir_cache(d_inode(dentry), NULL);
+		if (ovl_dir_cache(inode) == cache)
+			ovl_set_dir_cache(inode, NULL);
 
 		ovl_cache_free(&cache->entries);
 		kfree(cache);
@@ -323,15 +323,15 @@ static void ovl_dir_reset(struct file *file)
 {
 	struct ovl_dir_file *od = file->private_data;
 	struct ovl_dir_cache *cache = od->cache;
-	struct dentry *dentry = file->f_path.dentry;
+	struct inode *inode = file_inode(file);
 	bool is_real;
 
-	if (cache && ovl_dentry_version_get(dentry) != cache->version) {
-		ovl_cache_put(od, dentry);
+	if (cache && ovl_inode_version_get(inode) != cache->version) {
+		ovl_cache_put(od, inode);
 		od->cache = NULL;
 		od->cursor = NULL;
 	}
-	is_real = ovl_dir_is_real(dentry);
+	is_real = ovl_dir_is_real(inode);
 	if (od->is_real != is_real) {
 		/* is_real can only become false when dir is copied up */
 		if (WARN_ON(is_real))
@@ -394,9 +394,10 @@ static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
 {
 	int res;
 	struct ovl_dir_cache *cache;
+	struct inode *inode = d_inode(dentry);
 
-	cache = ovl_dir_cache(d_inode(dentry));
-	if (cache && ovl_dentry_version_get(dentry) == cache->version) {
+	cache = ovl_dir_cache(inode);
+	if (cache && ovl_inode_version_get(inode) == cache->version) {
 		WARN_ON(!cache->refcount);
 		cache->refcount++;
 		return cache;
@@ -418,8 +419,8 @@ static struct ovl_dir_cache *ovl_cache_get(struct dentry *dentry)
 		return ERR_PTR(res);
 	}
 
-	cache->version = ovl_dentry_version_get(dentry);
-	ovl_set_dir_cache(d_inode(dentry), cache);
+	cache->version = ovl_inode_version_get(inode);
+	ovl_set_dir_cache(inode, cache);
 
 	return cache;
 }
@@ -596,16 +597,17 @@ static struct ovl_dir_cache *ovl_cache_get_impure(const struct path *path)
 {
 	int res;
 	struct dentry *dentry = path->dentry;
+	struct inode *inode = d_inode(dentry);
 	struct ovl_fs *ofs = OVL_FS(dentry->d_sb);
 	struct ovl_dir_cache *cache;
 
-	cache = ovl_dir_cache(d_inode(dentry));
-	if (cache && ovl_dentry_version_get(dentry) == cache->version)
+	cache = ovl_dir_cache(inode);
+	if (cache && ovl_inode_version_get(inode) == cache->version)
 		return cache;
 
 	/* Impure cache is not refcounted, free it here */
-	ovl_dir_cache_free(d_inode(dentry));
-	ovl_set_dir_cache(d_inode(dentry), NULL);
+	ovl_dir_cache_free(inode);
+	ovl_set_dir_cache(inode, NULL);
 
 	cache = kzalloc(sizeof(struct ovl_dir_cache), GFP_KERNEL);
 	if (!cache)
@@ -627,13 +629,13 @@ static struct ovl_dir_cache *ovl_cache_get_impure(const struct path *path)
 					OVL_XATTR_IMPURE);
 			ovl_drop_write(dentry);
 		}
-		ovl_clear_flag(OVL_IMPURE, d_inode(dentry));
+		ovl_clear_flag(OVL_IMPURE, inode);
 		kfree(cache);
 		return NULL;
 	}
 
-	cache->version = ovl_dentry_version_get(dentry);
-	ovl_set_dir_cache(d_inode(dentry), cache);
+	cache->version = ovl_inode_version_get(inode);
+	ovl_set_dir_cache(inode, cache);
 
 	return cache;
 }
@@ -675,7 +677,7 @@ static bool ovl_fill_real(struct dir_context *ctx, const char *name,
 static bool ovl_is_impure_dir(struct file *file)
 {
 	struct ovl_dir_file *od = file->private_data;
-	struct inode *dir = d_inode(file->f_path.dentry);
+	struct inode *dir = file_inode(file);
 
 	/*
 	 * Only upper dir can be impure, but if we are in the middle of
@@ -893,7 +895,7 @@ static int ovl_dir_fsync(struct file *file, loff_t start, loff_t end,
 	struct file *realfile;
 	int err;
 
-	err = ovl_sync_status(OVL_FS(file->f_path.dentry->d_sb));
+	err = ovl_sync_status(OVL_FS(file_inode(file)->i_sb));
 	if (err <= 0)
 		return err;
 
@@ -913,7 +915,7 @@ static int ovl_dir_release(struct inode *inode, struct file *file)
 
 	if (od->cache) {
 		inode_lock(inode);
-		ovl_cache_put(od, file->f_path.dentry);
+		ovl_cache_put(od, inode);
 		inode_unlock(inode);
 	}
 	fput(od->realfile);
@@ -942,7 +944,7 @@ static int ovl_dir_open(struct inode *inode, struct file *file)
 		return PTR_ERR(realfile);
 	}
 	od->realfile = realfile;
-	od->is_real = ovl_dir_is_real(file->f_path.dentry);
+	od->is_real = ovl_dir_is_real(inode);
 	od->is_upper = OVL_TYPE_UPPER(type);
 	file->private_data = od;
 
@@ -1071,14 +1073,10 @@ static int ovl_workdir_cleanup_recurse(struct ovl_fs *ofs, const struct path *pa
 	int err;
 	struct inode *dir = path->dentry->d_inode;
 	LIST_HEAD(list);
-	struct rb_root root = RB_ROOT;
 	struct ovl_cache_entry *p;
 	struct ovl_readdir_data rdd = {
-		.ctx.actor = ovl_fill_merge,
-		.dentry = NULL,
+		.ctx.actor = ovl_fill_plain,
 		.list = &list,
-		.root = &root,
-		.is_lowest = false,
 	};
 	bool incompat = false;
 
@@ -1159,14 +1157,10 @@ int ovl_indexdir_cleanup(struct ovl_fs *ofs)
 	struct inode *dir = indexdir->d_inode;
 	struct path path = { .mnt = ovl_upper_mnt(ofs), .dentry = indexdir };
 	LIST_HEAD(list);
-	struct rb_root root = RB_ROOT;
 	struct ovl_cache_entry *p;
 	struct ovl_readdir_data rdd = {
-		.ctx.actor = ovl_fill_merge,
-		.dentry = NULL,
+		.ctx.actor = ovl_fill_plain,
 		.list = &list,
-		.root = &root,
-		.is_lowest = false,
 	};
 
 	err = ovl_dir_read(&path, &rdd);
diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
index 2addafe4e14a46..85b891152a2ca0 100644
--- a/fs/overlayfs/super.c
+++ b/fs/overlayfs/super.c
@@ -139,11 +139,16 @@ static int ovl_dentry_revalidate_common(struct dentry *dentry,
 					unsigned int flags, bool weak)
 {
 	struct ovl_entry *oe = dentry->d_fsdata;
+	struct inode *inode = d_inode_rcu(dentry);
 	struct dentry *upper;
 	unsigned int i;
 	int ret = 1;
 
-	upper = ovl_dentry_upper(dentry);
+	/* Careful in RCU mode */
+	if (!inode)
+		return -ECHILD;
+
+	upper = ovl_i_dentry_upper(inode);
 	if (upper)
 		ret = ovl_revalidate_real(upper, flags, weak);
 
diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c
index c0c20d33691b1e..bde291623c8ccf 100644
--- a/fs/overlayfs/util.c
+++ b/fs/overlayfs/util.c
@@ -463,7 +463,7 @@ static void ovl_dir_version_inc(struct dentry *dentry, bool impurity)
 	 * which have been copied up and have origins), so only need to note
 	 * changes to impure entries.
 	 */
-	if (!ovl_dir_is_real(dentry) || impurity)
+	if (!ovl_dir_is_real(inode) || impurity)
 		OVL_I(inode)->version++;
 }
 
@@ -475,10 +475,8 @@ void ovl_dir_modified(struct dentry *dentry, bool impurity)
 	ovl_dir_version_inc(dentry, impurity);
 }
 
-u64 ovl_dentry_version_get(struct dentry *dentry)
+u64 ovl_inode_version_get(struct inode *inode)
 {
-	struct inode *inode = d_inode(dentry);
-
 	WARN_ON(!inode_is_locked(inode));
 	return OVL_I(inode)->version;
 }