Merge tag 'fsnotify_for_v5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull fsnotify updates from Jan Kara:

 - fanotify fix for softlockups when there are many queued events

 - performance improvement to reduce fsnotify overhead when not used

 - Amir's implementation of fanotify events with names. With these you
   can now efficiently monitor whole filesystem, eg to mirror changes to
   another machine.

* tag 'fsnotify_for_v5.9-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs: (37 commits)
  fanotify: compare fsid when merging name event
  fsnotify: create method handle_inode_event() in fsnotify_operations
  fanotify: report parent fid + child fid
  fanotify: report parent fid + name + child fid
  fanotify: add support for FAN_REPORT_NAME
  fanotify: report events with parent dir fid to sb/mount/non-dir marks
  fanotify: add basic support for FAN_REPORT_DIR_FID
  fsnotify: remove check that source dentry is positive
  fsnotify: send event with parent/name info to sb/mount/non-dir marks
  audit: do not set FS_EVENT_ON_CHILD in audit marks mask
  inotify: do not set FS_EVENT_ON_CHILD in non-dir mark mask
  fsnotify: pass dir and inode arguments to fsnotify()
  fsnotify: create helper fsnotify_inode()
  fsnotify: send event to parent and child with single callback
  inotify: report both events on parent and child with single callback
  dnotify: report both events on parent and child with single callback
  fanotify: no external fh buffer in fanotify_name_event
  fanotify: use struct fanotify_info to parcel the variable size buffer
  fsnotify: add object type "child" to object type iterator
  fanotify: use FAN_EVENT_ON_CHILD as implicit flag on sb/mount/non-dir marks
  ...

4 files changed, 127 insertions, 72 deletions

diff --git a/include/linux/fanotify.h b/include/linux/fanotify.h
index b79fa9bb7359..3e9c56ee651f 100644
--- a/include/linux/fanotify.h
+++ b/include/linux/fanotify.h
@@ -18,8 +18,10 @@
 #define FANOTIFY_CLASS_BITS	(FAN_CLASS_NOTIF | FAN_CLASS_CONTENT | \
 				 FAN_CLASS_PRE_CONTENT)
 
-#define FANOTIFY_INIT_FLAGS	(FANOTIFY_CLASS_BITS | \
-				 FAN_REPORT_TID | FAN_REPORT_FID | \
+#define FANOTIFY_FID_BITS	(FAN_REPORT_FID | FAN_REPORT_DFID_NAME)
+
+#define FANOTIFY_INIT_FLAGS	(FANOTIFY_CLASS_BITS | FANOTIFY_FID_BITS | \
+				 FAN_REPORT_TID | \
 				 FAN_CLOEXEC | FAN_NONBLOCK | \
 				 FAN_UNLIMITED_QUEUE | FAN_UNLIMITED_MARKS)
 
diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h
index 5ab28f6c7d26..f8acddcf54fb 100644
--- a/include/linux/fsnotify.h
+++ b/include/linux/fsnotify.h
@@ -23,19 +23,14 @@
  * have changed (i.e. renamed over).
  *
  * Unlike fsnotify_parent(), the event will be reported regardless of the
- * FS_EVENT_ON_CHILD mask on the parent inode.
+ * FS_EVENT_ON_CHILD mask on the parent inode and will not be reported if only
+ * the child is interested and not the parent.
  */
 static inline void fsnotify_name(struct inode *dir, __u32 mask,
 				 struct inode *child,
 				 const struct qstr *name, u32 cookie)
 {
-	fsnotify(dir, mask, child, FSNOTIFY_EVENT_INODE, name, cookie);
-	/*
-	 * Send another flavor of the event without child inode data and
-	 * without the specific event type (e.g. FS_CREATE|FS_IS_DIR).
-	 * The name is relative to the dir inode the event is reported to.
-	 */
-	fsnotify(dir, FS_DIR_MODIFY, dir, FSNOTIFY_EVENT_INODE, name, 0);
+	fsnotify(mask, child, FSNOTIFY_EVENT_INODE, dir, name, NULL, cookie);
 }
 
 static inline void fsnotify_dirent(struct inode *dir, struct dentry *dentry,
@@ -44,38 +39,55 @@ static inline void fsnotify_dirent(struct inode *dir, struct dentry *dentry,
 	fsnotify_name(dir, mask, d_inode(dentry), &dentry->d_name, 0);
 }
 
-/*
- * Simple wrappers to consolidate calls fsnotify_parent()/fsnotify() when
- * an event is on a file/dentry.
- */
-static inline void fsnotify_dentry(struct dentry *dentry, __u32 mask)
+static inline void fsnotify_inode(struct inode *inode, __u32 mask)
+{
+	if (S_ISDIR(inode->i_mode))
+		mask |= FS_ISDIR;
+
+	fsnotify(mask, inode, FSNOTIFY_EVENT_INODE, NULL, NULL, inode, 0);
+}
+
+/* Notify this dentry's parent about a child's events. */
+static inline int fsnotify_parent(struct dentry *dentry, __u32 mask,
+				  const void *data, int data_type)
 {
 	struct inode *inode = d_inode(dentry);
 
-	if (S_ISDIR(inode->i_mode))
+	if (S_ISDIR(inode->i_mode)) {
 		mask |= FS_ISDIR;
 
-	fsnotify_parent(dentry, mask, inode, FSNOTIFY_EVENT_INODE);
-	fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+		/* sb/mount marks are not interested in name of directory */
+		if (!(dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED))
+			goto notify_child;
+	}
+
+	/* disconnected dentry cannot notify parent */
+	if (IS_ROOT(dentry))
+		goto notify_child;
+
+	return __fsnotify_parent(dentry, mask, data, data_type);
+
+notify_child:
+	return fsnotify(mask, data, data_type, NULL, NULL, inode, 0);
+}
+
+/*
+ * Simple wrappers to consolidate calls to fsnotify_parent() when an event
+ * is on a file/dentry.
+ */
+static inline void fsnotify_dentry(struct dentry *dentry, __u32 mask)
+{
+	fsnotify_parent(dentry, mask, d_inode(dentry), FSNOTIFY_EVENT_INODE);
 }
 
 static inline int fsnotify_file(struct file *file, __u32 mask)
 {
 	const struct path *path = &file->f_path;
-	struct inode *inode = file_inode(file);
-	int ret;
 
 	if (file->f_mode & FMODE_NONOTIFY)
 		return 0;
 
-	if (S_ISDIR(inode->i_mode))
-		mask |= FS_ISDIR;
-
-	ret = fsnotify_parent(path->dentry, mask, path, FSNOTIFY_EVENT_PATH);
-	if (ret)
-		return ret;
-
-	return fsnotify(inode, mask, path, FSNOTIFY_EVENT_PATH, NULL, 0);
+	return fsnotify_parent(path->dentry, mask, path, FSNOTIFY_EVENT_PATH);
 }
 
 /* Simple call site for access decisions */
@@ -108,12 +120,7 @@ static inline int fsnotify_perm(struct file *file, int mask)
  */
 static inline void fsnotify_link_count(struct inode *inode)
 {
-	__u32 mask = FS_ATTRIB;
-
-	if (S_ISDIR(inode->i_mode))
-		mask |= FS_ISDIR;
-
-	fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+	fsnotify_inode(inode, FS_ATTRIB);
 }
 
 /*
@@ -128,7 +135,6 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir,
 	u32 fs_cookie = fsnotify_get_cookie();
 	__u32 old_dir_mask = FS_MOVED_FROM;
 	__u32 new_dir_mask = FS_MOVED_TO;
-	__u32 mask = FS_MOVE_SELF;
 	const struct qstr *new_name = &moved->d_name;
 
 	if (old_dir == new_dir)
@@ -137,7 +143,6 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir,
 	if (isdir) {
 		old_dir_mask |= FS_ISDIR;
 		new_dir_mask |= FS_ISDIR;
-		mask |= FS_ISDIR;
 	}
 
 	fsnotify_name(old_dir, old_dir_mask, source, old_name, fs_cookie);
@@ -145,9 +150,7 @@ static inline void fsnotify_move(struct inode *old_dir, struct inode *new_dir,
 
 	if (target)
 		fsnotify_link_count(target);
-
-	if (source)
-		fsnotify(source, mask, source, FSNOTIFY_EVENT_INODE, NULL, 0);
+	fsnotify_inode(source, FS_MOVE_SELF);
 	audit_inode_child(new_dir, moved, AUDIT_TYPE_CHILD_CREATE);
 }
 
@@ -172,12 +175,7 @@ static inline void fsnotify_vfsmount_delete(struct vfsmount *mnt)
  */
 static inline void fsnotify_inoderemove(struct inode *inode)
 {
-	__u32 mask = FS_DELETE_SELF;
-
-	if (S_ISDIR(inode->i_mode))
-		mask |= FS_ISDIR;
-
-	fsnotify(inode, mask, inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+	fsnotify_inode(inode, FS_DELETE_SELF);
 	__fsnotify_inode_delete(inode);
 }
 
diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h
index f0c506405b54..f8529a3a2923 100644
--- a/include/linux/fsnotify_backend.h
+++ b/include/linux/fsnotify_backend.h
@@ -47,11 +47,13 @@
 #define FS_OPEN_PERM		0x00010000	/* open event in an permission hook */
 #define FS_ACCESS_PERM		0x00020000	/* access event in a permissions hook */
 #define FS_OPEN_EXEC_PERM	0x00040000	/* open/exec event in a permission hook */
-#define FS_DIR_MODIFY		0x00080000	/* Directory entry was modified */
 
 #define FS_EXCL_UNLINK		0x04000000	/* do not send events if object is unlinked */
-/* This inode cares about things that happen to its children.  Always set for
- * dnotify and inotify. */
+/*
+ * Set on inode mark that cares about things that happen to its children.
+ * Always set for dnotify and inotify.
+ * Set on inode/sb/mount marks that care about parent/name info.
+ */
 #define FS_EVENT_ON_CHILD	0x08000000
 
 #define FS_DN_RENAME		0x10000000	/* file renamed */
@@ -67,21 +69,28 @@
  * The watching parent may get an FS_ATTRIB|FS_EVENT_ON_CHILD event
  * when a directory entry inside a child subdir changes.
  */
-#define ALL_FSNOTIFY_DIRENT_EVENTS	(FS_CREATE | FS_DELETE | FS_MOVE | \
-					 FS_DIR_MODIFY)
+#define ALL_FSNOTIFY_DIRENT_EVENTS	(FS_CREATE | FS_DELETE | FS_MOVE)
 
 #define ALL_FSNOTIFY_PERM_EVENTS (FS_OPEN_PERM | FS_ACCESS_PERM | \
 				  FS_OPEN_EXEC_PERM)
 
 /*
- * This is a list of all events that may get sent to a parent based on fs event
- * happening to inodes inside that directory.
+ * This is a list of all events that may get sent to a parent that is watching
+ * with flag FS_EVENT_ON_CHILD based on fs event on a child of that directory.
  */
 #define FS_EVENTS_POSS_ON_CHILD   (ALL_FSNOTIFY_PERM_EVENTS | \
 				   FS_ACCESS | FS_MODIFY | FS_ATTRIB | \
 				   FS_CLOSE_WRITE | FS_CLOSE_NOWRITE | \
 				   FS_OPEN | FS_OPEN_EXEC)
 
+/*
+ * This is a list of all events that may get sent with the parent inode as the
+ * @to_tell argument of fsnotify().
+ * It may include events that can be sent to an inode/sb/mount mark, but cannot
+ * be sent to a parent watching children.
+ */
+#define FS_EVENTS_POSS_TO_PARENT (FS_EVENTS_POSS_ON_CHILD)
+
 /* Events that can be reported to backends */
 #define ALL_FSNOTIFY_EVENTS (ALL_FSNOTIFY_DIRENT_EVENTS | \
 			     FS_EVENTS_POSS_ON_CHILD | \
@@ -108,18 +117,41 @@ struct mem_cgroup;
  * these operations for each relevant group.
  *
  * handle_event - main call for a group to handle an fs event
+ * @group:	group to notify
+ * @mask:	event type and flags
+ * @data:	object that event happened on
+ * @data_type:	type of object for fanotify_data_XXX() accessors
+ * @dir:	optional directory associated with event -
+ *		if @file_name is not NULL, this is the directory that
+ *		@file_name is relative to
+ * @file_name:	optional file name associated with event
+ * @cookie:	inotify rename cookie
+ * @iter_info:	array of marks from this group that are interested in the event
+ *
+ * handle_inode_event - simple variant of handle_event() for groups that only
+ *		have inode marks and don't have ignore mask
+ * @mark:	mark to notify
+ * @mask:	event type and flags
+ * @inode:	inode that event happened on
+ * @dir:	optional directory associated with event -
+ *		if @file_name is not NULL, this is the directory that
+ *		@file_name is relative to.
+ * @file_name:	optional file name associated with event
+ *
  * free_group_priv - called when a group refcnt hits 0 to clean up the private union
  * freeing_mark - called when a mark is being destroyed for some reason.  The group
- * 		MUST be holding a reference on each mark and that reference must be
- * 		dropped in this function.  inotify uses this function to send
- * 		userspace messages that marks have been removed.
+ *		MUST be holding a reference on each mark and that reference must be
+ *		dropped in this function.  inotify uses this function to send
+ *		userspace messages that marks have been removed.
  */
 struct fsnotify_ops {
-	int (*handle_event)(struct fsnotify_group *group,
-			    struct inode *inode,
-			    u32 mask, const void *data, int data_type,
+	int (*handle_event)(struct fsnotify_group *group, u32 mask,
+			    const void *data, int data_type, struct inode *dir,
 			    const struct qstr *file_name, u32 cookie,
 			    struct fsnotify_iter_info *iter_info);
+	int (*handle_inode_event)(struct fsnotify_mark *mark, u32 mask,
+			    struct inode *inode, struct inode *dir,
+			    const struct qstr *file_name);
 	void (*free_group_priv)(struct fsnotify_group *group);
 	void (*freeing_mark)(struct fsnotify_mark *mark, struct fsnotify_group *group);
 	void (*free_event)(struct fsnotify_event *event);
@@ -220,12 +252,11 @@ enum fsnotify_data_type {
 	FSNOTIFY_EVENT_INODE,
 };
 
-static inline const struct inode *fsnotify_data_inode(const void *data,
-						      int data_type)
+static inline struct inode *fsnotify_data_inode(const void *data, int data_type)
 {
 	switch (data_type) {
 	case FSNOTIFY_EVENT_INODE:
-		return data;
+		return (struct inode *)data;
 	case FSNOTIFY_EVENT_PATH:
 		return d_inode(((const struct path *)data)->dentry);
 	default:
@@ -246,6 +277,7 @@ static inline const struct path *fsnotify_data_path(const void *data,
 
 enum fsnotify_obj_type {
 	FSNOTIFY_OBJ_TYPE_INODE,
+	FSNOTIFY_OBJ_TYPE_CHILD,
 	FSNOTIFY_OBJ_TYPE_VFSMOUNT,
 	FSNOTIFY_OBJ_TYPE_SB,
 	FSNOTIFY_OBJ_TYPE_COUNT,
@@ -253,6 +285,7 @@ enum fsnotify_obj_type {
 };
 
 #define FSNOTIFY_OBJ_TYPE_INODE_FL	(1U << FSNOTIFY_OBJ_TYPE_INODE)
+#define FSNOTIFY_OBJ_TYPE_CHILD_FL	(1U << FSNOTIFY_OBJ_TYPE_CHILD)
 #define FSNOTIFY_OBJ_TYPE_VFSMOUNT_FL	(1U << FSNOTIFY_OBJ_TYPE_VFSMOUNT)
 #define FSNOTIFY_OBJ_TYPE_SB_FL		(1U << FSNOTIFY_OBJ_TYPE_SB)
 #define FSNOTIFY_OBJ_ALL_TYPES_MASK	((1U << FSNOTIFY_OBJ_TYPE_COUNT) - 1)
@@ -297,6 +330,7 @@ static inline struct fsnotify_mark *fsnotify_iter_##name##_mark( \
 }
 
 FSNOTIFY_ITER_FUNCS(inode, INODE)
+FSNOTIFY_ITER_FUNCS(child, CHILD)
 FSNOTIFY_ITER_FUNCS(vfsmount, VFSMOUNT)
 FSNOTIFY_ITER_FUNCS(sb, SB)
 
@@ -377,15 +411,29 @@ struct fsnotify_mark {
 /* called from the vfs helpers */
 
 /* main fsnotify call to send events */
-extern int fsnotify(struct inode *to_tell, __u32 mask, const void *data,
-		    int data_type, const struct qstr *name, u32 cookie);
-extern int fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
+extern int fsnotify(__u32 mask, const void *data, int data_type,
+		    struct inode *dir, const struct qstr *name,
+		    struct inode *inode, u32 cookie);
+extern int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
 			   int data_type);
 extern void __fsnotify_inode_delete(struct inode *inode);
 extern void __fsnotify_vfsmount_delete(struct vfsmount *mnt);
 extern void fsnotify_sb_delete(struct super_block *sb);
 extern u32 fsnotify_get_cookie(void);
 
+static inline __u32 fsnotify_parent_needed_mask(__u32 mask)
+{
+	/* FS_EVENT_ON_CHILD is set on marks that want parent/name info */
+	if (!(mask & FS_EVENT_ON_CHILD))
+		return 0;
+	/*
+	 * This object might be watched by a mark that cares about parent/name
+	 * info, does it care about the specific set of events that can be
+	 * reported with parent/name info?
+	 */
+	return mask & FS_EVENTS_POSS_TO_PARENT;
+}
+
 static inline int fsnotify_inode_watches_children(struct inode *inode)
 {
 	/* FS_EVENT_ON_CHILD is set if the inode may care */
@@ -535,13 +583,14 @@ static inline void fsnotify_init_event(struct fsnotify_event *event,
 
 #else
 
-static inline int fsnotify(struct inode *to_tell, __u32 mask, const void *data,
-			   int data_type, const struct qstr *name, u32 cookie)
+static inline int fsnotify(__u32 mask, const void *data, int data_type,
+			   struct inode *dir, const struct qstr *name,
+			   struct inode *inode, u32 cookie)
 {
 	return 0;
 }
 
-static inline int fsnotify_parent(struct dentry *dentry, __u32 mask,
+static inline int __fsnotify_parent(struct dentry *dentry, __u32 mask,
 				  const void *data, int data_type)
 {
 	return 0;
diff --git a/include/uapi/linux/fanotify.h b/include/uapi/linux/fanotify.h
index a88c7c6d0692..fbf9c5c7dd59 100644
--- a/include/uapi/linux/fanotify.h
+++ b/include/uapi/linux/fanotify.h
@@ -24,7 +24,6 @@
 #define FAN_OPEN_PERM		0x00010000	/* File open in perm check */
 #define FAN_ACCESS_PERM		0x00020000	/* File accessed in perm check */
 #define FAN_OPEN_EXEC_PERM	0x00040000	/* File open/exec in perm check */
-#define FAN_DIR_MODIFY		0x00080000	/* Directory entry was modified */
 
 #define FAN_EVENT_ON_CHILD	0x08000000	/* Interested in child events */
 
@@ -54,6 +53,11 @@
 /* Flags to determine fanotify event format */
 #define FAN_REPORT_TID		0x00000100	/* event->pid is thread id */
 #define FAN_REPORT_FID		0x00000200	/* Report unique file id */
+#define FAN_REPORT_DIR_FID	0x00000400	/* Report unique directory id */
+#define FAN_REPORT_NAME		0x00000800	/* Report events with name */
+
+/* Convenience macro - FAN_REPORT_NAME requires FAN_REPORT_DIR_FID */
+#define FAN_REPORT_DFID_NAME	(FAN_REPORT_DIR_FID | FAN_REPORT_NAME)
 
 /* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_INIT_FLAGS	(FAN_CLOEXEC | FAN_NONBLOCK | \
@@ -118,6 +122,7 @@ struct fanotify_event_metadata {
 
 #define FAN_EVENT_INFO_TYPE_FID		1
 #define FAN_EVENT_INFO_TYPE_DFID_NAME	2
+#define FAN_EVENT_INFO_TYPE_DFID	3
 
 /* Variable length info record following event metadata */
 struct fanotify_event_info_header {
@@ -127,10 +132,11 @@ struct fanotify_event_info_header {
 };
 
 /*
- * Unique file identifier info record. This is used both for
- * FAN_EVENT_INFO_TYPE_FID records and for FAN_EVENT_INFO_TYPE_DFID_NAME
- * records. For FAN_EVENT_INFO_TYPE_DFID_NAME there is additionally a null
- * terminated name immediately after the file handle.
+ * Unique file identifier info record.
+ * This structure is used for records of types FAN_EVENT_INFO_TYPE_FID,
+ * FAN_EVENT_INFO_TYPE_DFID and FAN_EVENT_INFO_TYPE_DFID_NAME.
+ * For FAN_EVENT_INFO_TYPE_DFID_NAME there is additionally a null terminated
+ * name immediately after the file handle.
  */
 struct fanotify_event_info_fid {
 	struct fanotify_event_info_header hdr;