Merge tag 'for_v4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull fsnotify updates from Jan Kara:
 "Amir's patches to implement superblock fanotify watches, Xiaoming's
  patch to enable reporting of thread IDs in fanotify events instead of
  TGIDs (sadly the patch got mis-attributed to Amir and I've noticed
  only now), and a fix of possible oops on umount caused by fsnotify
  infrastructure"

* tag 'for_v4.20-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  fsnotify: Fix busy inodes during unmount
  fs: group frequently accessed fields of struct super_block together
  fanotify: support reporting thread id instead of process id
  fanotify: add BUILD_BUG_ON() to count the bits of fanotify constants
  fsnotify: convert runtime BUG_ON() to BUILD_BUG_ON()
  fanotify: deprecate uapi FAN_ALL_* constants
  fanotify: simplify handling of FAN_ONDIR
  fsnotify: generalize handling of extra event flags
  fanotify: fix collision of internal and uapi mark flags
  fanotify: store fanotify_init() flags in group's fanotify_data
  fanotify: add API to attach/detach super block mark
  fsnotify: send path type events to group with super block marks
  fsnotify: add super block object type

fs/notify/fanotify/fanotify.c

@@ -25,7 +25,7 @@ static bool should_merge(struct fsnotify_event *old_fsn,
 	old = FANOTIFY_E(old_fsn);
 	new = FANOTIFY_E(new_fsn);
 
-	if (old_fsn->inode == new_fsn->inode && old->tgid == new->tgid &&
+	if (old_fsn->inode == new_fsn->inode && old->pid == new->pid &&
 	    old->path.mnt == new->path.mnt &&
 	    old->path.dentry == new->path.dentry)
 		return true;
@@ -131,8 +131,8 @@ static bool fanotify_should_send_event(struct fsnotify_iter_info *iter_info,
 	    !(marks_mask & FS_ISDIR & ~marks_ignored_mask))
 		return false;
 
-	if (event_mask & FAN_ALL_OUTGOING_EVENTS & marks_mask &
-				 ~marks_ignored_mask)
+	if (event_mask & FANOTIFY_OUTGOING_EVENTS &
+	    marks_mask & ~marks_ignored_mask)
 		return true;
 
 	return false;
@@ -171,7 +171,10 @@ struct fanotify_event_info *fanotify_alloc_event(struct fsnotify_group *group,
 		goto out;
 init: __maybe_unused
 	fsnotify_init_event(&event->fse, inode, mask);
-	event->tgid = get_pid(task_tgid(current));
+	if (FAN_GROUP_FLAG(group, FAN_REPORT_TID))
+		event->pid = get_pid(task_pid(current));
+	else
+		event->pid = get_pid(task_tgid(current));
 	if (path) {
 		event->path = *path;
 		path_get(&event->path);
@@ -205,6 +208,8 @@ static int fanotify_handle_event(struct fsnotify_group *group,
 	BUILD_BUG_ON(FAN_ACCESS_PERM != FS_ACCESS_PERM);
 	BUILD_BUG_ON(FAN_ONDIR != FS_ISDIR);
 
+	BUILD_BUG_ON(HWEIGHT32(ALL_FANOTIFY_EVENT_BITS) != 10);
+
 	if (!fanotify_should_send_event(iter_info, mask, data, data_type))
 		return 0;
 
@@ -236,7 +241,7 @@ static int fanotify_handle_event(struct fsnotify_group *group,
 	ret = fsnotify_add_event(group, fsn_event, fanotify_merge);
 	if (ret) {
 		/* Permission events shouldn't be merged */
-		BUG_ON(ret == 1 && mask & FAN_ALL_PERM_EVENTS);
+		BUG_ON(ret == 1 && mask & FANOTIFY_PERM_EVENTS);
 		/* Our event wasn't used in the end. Free it. */
 		fsnotify_destroy_event(group, fsn_event);
 
@@ -268,7 +273,7 @@ static void fanotify_free_event(struct fsnotify_event *fsn_event)
 
 	event = FANOTIFY_E(fsn_event);
 	path_put(&event->path);
-	put_pid(event->tgid);
+	put_pid(event->pid);
 	if (fanotify_is_perm_event(fsn_event->mask)) {
 		kmem_cache_free(fanotify_perm_event_cachep,
 				FANOTIFY_PE(fsn_event));

fs/notify/fanotify/fanotify.h

@@ -19,7 +19,7 @@ struct fanotify_event_info {
 	 * during this object's lifetime
 	 */
 	struct path path;
-	struct pid *tgid;
+	struct pid *pid;
 };
 
 /*
@@ -44,7 +44,7 @@ FANOTIFY_PE(struct fsnotify_event *fse)
 static inline bool fanotify_is_perm_event(u32 mask)
 {
 	return IS_ENABLED(CONFIG_FANOTIFY_ACCESS_PERMISSIONS) &&
-		mask & FAN_ALL_PERM_EVENTS;
+		mask & FANOTIFY_PERM_EVENTS;
 }
 
 static inline struct fanotify_event_info *FANOTIFY_E(struct fsnotify_event *fse)

fs/notify/fanotify/fanotify_user.c

@@ -131,8 +131,8 @@ static int fill_event_metadata(struct fsnotify_group *group,
 	metadata->metadata_len = FAN_EVENT_METADATA_LEN;
 	metadata->vers = FANOTIFY_METADATA_VERSION;
 	metadata->reserved = 0;
-	metadata->mask = fsn_event->mask & FAN_ALL_OUTGOING_EVENTS;
-	metadata->pid = pid_vnr(event->tgid);
+	metadata->mask = fsn_event->mask & FANOTIFY_OUTGOING_EVENTS;
+	metadata->pid = pid_vnr(event->pid);
 	if (unlikely(fsn_event->mask & FAN_Q_OVERFLOW))
 		metadata->fd = FAN_NOFD;
 	else {
@@ -191,7 +191,7 @@ static int process_access_response(struct fsnotify_group *group,
 	if (fd < 0)
 		return -EINVAL;
 
-	if ((response & FAN_AUDIT) && !group->fanotify_data.audit)
+	if ((response & FAN_AUDIT) && !FAN_GROUP_FLAG(group, FAN_ENABLE_AUDIT))
 		return -EINVAL;
 
 	event = dequeue_event(group, fd);
@@ -395,7 +395,7 @@ static int fanotify_release(struct inode *ignored, struct file *file)
 	 */
 	while (!fsnotify_notify_queue_is_empty(group)) {
 		fsn_event = fsnotify_remove_first_event(group);
-		if (!(fsn_event->mask & FAN_ALL_PERM_EVENTS)) {
+		if (!(fsn_event->mask & FANOTIFY_PERM_EVENTS)) {
 			spin_unlock(&group->notification_lock);
 			fsnotify_destroy_event(group, fsn_event);
 			spin_lock(&group->notification_lock);
@@ -506,18 +506,10 @@ static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
 
 	spin_lock(&fsn_mark->lock);
 	if (!(flags & FAN_MARK_IGNORED_MASK)) {
-		__u32 tmask = fsn_mark->mask & ~mask;
-
-		if (flags & FAN_MARK_ONDIR)
-			tmask &= ~FAN_ONDIR;
-
 		oldmask = fsn_mark->mask;
-		fsn_mark->mask = tmask;
+		fsn_mark->mask &= ~mask;
 	} else {
-		__u32 tmask = fsn_mark->ignored_mask & ~mask;
-		if (flags & FAN_MARK_ONDIR)
-			tmask &= ~FAN_ONDIR;
-		fsn_mark->ignored_mask = tmask;
+		fsn_mark->ignored_mask &= ~mask;
 	}
 	*destroy = !(fsn_mark->mask | fsn_mark->ignored_mask);
 	spin_unlock(&fsn_mark->lock);
@@ -563,6 +555,13 @@ static int fanotify_remove_vfsmount_mark(struct fsnotify_group *group,
 				    mask, flags);
 }
 
+static int fanotify_remove_sb_mark(struct fsnotify_group *group,
+				      struct super_block *sb, __u32 mask,
+				      unsigned int flags)
+{
+	return fanotify_remove_mark(group, &sb->s_fsnotify_marks, mask, flags);
+}
+
 static int fanotify_remove_inode_mark(struct fsnotify_group *group,
 				      struct inode *inode, __u32 mask,
 				      unsigned int flags)
@@ -579,19 +578,10 @@ static __u32 fanotify_mark_add_to_mask(struct fsnotify_mark *fsn_mark,
 
 	spin_lock(&fsn_mark->lock);
 	if (!(flags & FAN_MARK_IGNORED_MASK)) {
-		__u32 tmask = fsn_mark->mask | mask;
-
-		if (flags & FAN_MARK_ONDIR)
-			tmask |= FAN_ONDIR;
-
 		oldmask = fsn_mark->mask;
-		fsn_mark->mask = tmask;
+		fsn_mark->mask |= mask;
 	} else {
-		__u32 tmask = fsn_mark->ignored_mask | mask;
-		if (flags & FAN_MARK_ONDIR)
-			tmask |= FAN_ONDIR;
-
-		fsn_mark->ignored_mask = tmask;
+		fsn_mark->ignored_mask |= mask;
 		if (flags & FAN_MARK_IGNORED_SURV_MODIFY)
 			fsn_mark->flags |= FSNOTIFY_MARK_FLAG_IGNORED_SURV_MODIFY;
 	}
@@ -658,6 +648,14 @@ static int fanotify_add_vfsmount_mark(struct fsnotify_group *group,
 				 FSNOTIFY_OBJ_TYPE_VFSMOUNT, mask, flags);
 }
 
+static int fanotify_add_sb_mark(struct fsnotify_group *group,
+				      struct super_block *sb, __u32 mask,
+				      unsigned int flags)
+{
+	return fanotify_add_mark(group, &sb->s_fsnotify_marks,
+				 FSNOTIFY_OBJ_TYPE_SB, mask, flags);
+}
+
 static int fanotify_add_inode_mark(struct fsnotify_group *group,
 				   struct inode *inode, __u32 mask,
 				   unsigned int flags)
@@ -686,16 +684,16 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	struct user_struct *user;
 	struct fanotify_event_info *oevent;
 
-	pr_debug("%s: flags=%d event_f_flags=%d\n",
+	pr_debug("%s: flags=%x event_f_flags=%x\n",
 		 __func__, flags, event_f_flags);
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
 #ifdef CONFIG_AUDITSYSCALL
-	if (flags & ~(FAN_ALL_INIT_FLAGS | FAN_ENABLE_AUDIT))
+	if (flags & ~(FANOTIFY_INIT_FLAGS | FAN_ENABLE_AUDIT))
 #else
-	if (flags & ~FAN_ALL_INIT_FLAGS)
+	if (flags & ~FANOTIFY_INIT_FLAGS)
 #endif
 		return -EINVAL;
 
@@ -731,6 +729,7 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	}
 
 	group->fanotify_data.user = user;
+	group->fanotify_data.flags = flags;
 	atomic_inc(&user->fanotify_listeners);
 	group->memcg = get_mem_cgroup_from_mm(current->mm);
 
@@ -746,7 +745,7 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	group->fanotify_data.f_flags = event_f_flags;
 	init_waitqueue_head(&group->fanotify_data.access_waitq);
 	INIT_LIST_HEAD(&group->fanotify_data.access_list);
-	switch (flags & FAN_ALL_CLASS_BITS) {
+	switch (flags & FANOTIFY_CLASS_BITS) {
 	case FAN_CLASS_NOTIF:
 		group->priority = FS_PRIO_0;
 		break;
@@ -783,7 +782,6 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 		fd = -EPERM;
 		if (!capable(CAP_AUDIT_WRITE))
 			goto out_destroy_group;
-		group->fanotify_data.audit = true;
 	}
 
 	fd = anon_inode_getfd("[fanotify]", &fanotify_fops, group, f_flags);
@@ -805,7 +803,8 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	struct fsnotify_group *group;
 	struct fd f;
 	struct path path;
-	u32 valid_mask = FAN_ALL_EVENTS | FAN_EVENT_ON_CHILD;
+	u32 valid_mask = FANOTIFY_EVENTS | FANOTIFY_EVENT_FLAGS;
+	unsigned int mark_type = flags & FANOTIFY_MARK_TYPE_BITS;
 	int ret;
 
 	pr_debug("%s: fanotify_fd=%d flags=%x dfd=%d pathname=%p mask=%llx\n",
@@ -815,8 +814,18 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	if (mask & ((__u64)0xffffffff << 32))
 		return -EINVAL;
 
-	if (flags & ~FAN_ALL_MARK_FLAGS)
+	if (flags & ~FANOTIFY_MARK_FLAGS)
+		return -EINVAL;
+
+	switch (mark_type) {
+	case FAN_MARK_INODE:
+	case FAN_MARK_MOUNT:
+	case FAN_MARK_FILESYSTEM:
+		break;
+	default:
 		return -EINVAL;
+	}
+
 	switch (flags & (FAN_MARK_ADD | FAN_MARK_REMOVE | FAN_MARK_FLUSH)) {
 	case FAN_MARK_ADD:		/* fallthrough */
 	case FAN_MARK_REMOVE:
@@ -824,20 +833,15 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 			return -EINVAL;
 		break;
 	case FAN_MARK_FLUSH:
-		if (flags & ~(FAN_MARK_MOUNT | FAN_MARK_FLUSH))
+		if (flags & ~(FANOTIFY_MARK_TYPE_BITS | FAN_MARK_FLUSH))
 			return -EINVAL;
 		break;
 	default:
 		return -EINVAL;
 	}
 
-	if (mask & FAN_ONDIR) {
-		flags |= FAN_MARK_ONDIR;
-		mask &= ~FAN_ONDIR;
-	}
-
 	if (IS_ENABLED(CONFIG_FANOTIFY_ACCESS_PERMISSIONS))
-		valid_mask |= FAN_ALL_PERM_EVENTS;
+		valid_mask |= FANOTIFY_PERM_EVENTS;
 
 	if (mask & ~valid_mask)
 		return -EINVAL;
@@ -857,14 +861,16 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	 * allowed to set permissions events.
 	 */
 	ret = -EINVAL;
-	if (mask & FAN_ALL_PERM_EVENTS &&
+	if (mask & FANOTIFY_PERM_EVENTS &&
 	    group->priority == FS_PRIO_0)
 		goto fput_and_out;
 
 	if (flags & FAN_MARK_FLUSH) {
 		ret = 0;
-		if (flags & FAN_MARK_MOUNT)
+		if (mark_type == FAN_MARK_MOUNT)
 			fsnotify_clear_vfsmount_marks_by_group(group);
+		else if (mark_type == FAN_MARK_FILESYSTEM)
+			fsnotify_clear_sb_marks_by_group(group);
 		else
 			fsnotify_clear_inode_marks_by_group(group);
 		goto fput_and_out;
@@ -875,7 +881,7 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 		goto fput_and_out;
 
 	/* inode held in place by reference to path; group by fget on fd */
-	if (!(flags & FAN_MARK_MOUNT))
+	if (mark_type == FAN_MARK_INODE)
 		inode = path.dentry->d_inode;
 	else
 		mnt = path.mnt;
@@ -883,14 +889,18 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	/* create/update an inode mark */
 	switch (flags & (FAN_MARK_ADD | FAN_MARK_REMOVE)) {
 	case FAN_MARK_ADD:
-		if (flags & FAN_MARK_MOUNT)
+		if (mark_type == FAN_MARK_MOUNT)
 			ret = fanotify_add_vfsmount_mark(group, mnt, mask, flags);
+		else if (mark_type == FAN_MARK_FILESYSTEM)
+			ret = fanotify_add_sb_mark(group, mnt->mnt_sb, mask, flags);
 		else
 			ret = fanotify_add_inode_mark(group, inode, mask, flags);
 		break;
 	case FAN_MARK_REMOVE:
-		if (flags & FAN_MARK_MOUNT)
+		if (mark_type == FAN_MARK_MOUNT)
 			ret = fanotify_remove_vfsmount_mark(group, mnt, mask, flags);
+		else if (mark_type == FAN_MARK_FILESYSTEM)
+			ret = fanotify_remove_sb_mark(group, mnt->mnt_sb, mask, flags);
 		else
 			ret = fanotify_remove_inode_mark(group, inode, mask, flags);
 		break;
@@ -934,6 +944,9 @@ COMPAT_SYSCALL_DEFINE6(fanotify_mark,
  */
 static int __init fanotify_user_setup(void)
 {
+	BUILD_BUG_ON(HWEIGHT32(FANOTIFY_INIT_FLAGS) != 7);
+	BUILD_BUG_ON(HWEIGHT32(FANOTIFY_MARK_FLAGS) != 9);
+
 	fanotify_mark_cache = KMEM_CACHE(fsnotify_mark,
 					 SLAB_PANIC|SLAB_ACCOUNT);
 	fanotify_event_cachep = KMEM_CACHE(fanotify_event_info, SLAB_PANIC);

fs/notify/fdinfo.c

@@ -131,37 +131,20 @@ static void fanotify_fdinfo(struct seq_file *m, struct fsnotify_mark *mark)
 
 		seq_printf(m, "fanotify mnt_id:%x mflags:%x mask:%x ignored_mask:%x\n",
 			   mnt->mnt_id, mflags, mark->mask, mark->ignored_mask);
+	} else if (mark->connector->type == FSNOTIFY_OBJ_TYPE_SB) {
+		struct super_block *sb = fsnotify_conn_sb(mark->connector);
+
+		seq_printf(m, "fanotify sdev:%x mflags:%x mask:%x ignored_mask:%x\n",
+			   sb->s_dev, mflags, mark->mask, mark->ignored_mask);
 	}
 }
 
 void fanotify_show_fdinfo(struct seq_file *m, struct file *f)
 {
 	struct fsnotify_group *group = f->private_data;
-	unsigned int flags = 0;
-
-	switch (group->priority) {
-	case FS_PRIO_0:
-		flags |= FAN_CLASS_NOTIF;
-		break;
-	case FS_PRIO_1:
-		flags |= FAN_CLASS_CONTENT;
-		break;
-	case FS_PRIO_2:
-		flags |= FAN_CLASS_PRE_CONTENT;
-		break;
-	}
-
-	if (group->max_events == UINT_MAX)
-		flags |= FAN_UNLIMITED_QUEUE;
-
-	if (group->fanotify_data.max_marks == UINT_MAX)
-		flags |= FAN_UNLIMITED_MARKS;
-
-	if (group->fanotify_data.audit)
-		flags |= FAN_ENABLE_AUDIT;
 
 	seq_printf(m, "fanotify flags:%x event-flags:%x\n",
-		   flags, group->fanotify_data.f_flags);
+		   group->fanotify_data.flags, group->fanotify_data.f_flags);
 
 	show_fdinfo(m, f, fanotify_fdinfo);
 }

fs/notify/fsnotify.c

@@ -48,7 +48,7 @@ void __fsnotify_vfsmount_delete(struct vfsmount *mnt)
  * Called during unmount with no locks held, so needs to be safe against
  * concurrent modifiers. We temporarily drop sb->s_inode_list_lock and CAN block.
  */
-void fsnotify_unmount_inodes(struct super_block *sb)
+static void fsnotify_unmount_inodes(struct super_block *sb)
 {
 	struct inode *inode, *iput_inode = NULL;
 
@@ -96,6 +96,15 @@ void fsnotify_unmount_inodes(struct super_block *sb)
 
 	if (iput_inode)
 		iput(iput_inode);
+	/* Wait for outstanding inode references from connectors */
+	wait_var_event(&sb->s_fsnotify_inode_refs,
+		       !atomic_long_read(&sb->s_fsnotify_inode_refs));
+}
+
+void fsnotify_sb_delete(struct super_block *sb)
+{
+	fsnotify_unmount_inodes(sb);
+	fsnotify_clear_marks_by_sb(sb);
 }
 
 /*
@@ -190,7 +199,7 @@ static int send_to_group(struct inode *to_tell,
 			 struct fsnotify_iter_info *iter_info)
 {
 	struct fsnotify_group *group = NULL;
-	__u32 test_mask = (mask & ~FS_EVENT_ON_CHILD);
+	__u32 test_mask = (mask & ALL_FSNOTIFY_EVENTS);
 	__u32 marks_mask = 0;
 	__u32 marks_ignored_mask = 0;
 	struct fsnotify_mark *mark;
@@ -319,15 +328,17 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
 	     const unsigned char *file_name, u32 cookie)
 {
 	struct fsnotify_iter_info iter_info = {};
-	struct mount *mnt;
+	struct super_block *sb = NULL;
+	struct mount *mnt = NULL;
+	__u32 mnt_or_sb_mask = 0;
 	int ret = 0;
-	/* global tests shouldn't care about events on child only the specific event */
-	__u32 test_mask = (mask & ~FS_EVENT_ON_CHILD);
+	__u32 test_mask = (mask & ALL_FSNOTIFY_EVENTS);
 
-	if (data_is == FSNOTIFY_EVENT_PATH)
+	if (data_is == FSNOTIFY_EVENT_PATH) {
 		mnt = real_mount(((const struct path *)data)->mnt);
-	else
-		mnt = NULL;
+		sb = mnt->mnt.mnt_sb;
+		mnt_or_sb_mask = mnt->mnt_fsnotify_mask | sb->s_fsnotify_mask;
+	}
 
 	/*
 	 * Optimization: srcu_read_lock() has a memory barrier which can
@@ -337,16 +348,15 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
 	 * need SRCU to keep them "alive".
 	 */
 	if (!to_tell->i_fsnotify_marks &&
-	    (!mnt || !mnt->mnt_fsnotify_marks))
+	    (!mnt || (!mnt->mnt_fsnotify_marks && !sb->s_fsnotify_marks)))
 		return 0;
 	/*
 	 * if this is a modify event we may need to clear the ignored masks
-	 * otherwise return if neither the inode nor the vfsmount care about
+	 * otherwise return if neither the inode nor the vfsmount/sb care about
 	 * this type of event.
 	 */
 	if (!(mask & FS_MODIFY) &&
-	    !(test_mask & to_tell->i_fsnotify_mask) &&
-	    !(mnt && test_mask & mnt->mnt_fsnotify_mask))
+	    !(test_mask & (to_tell->i_fsnotify_mask | mnt_or_sb_mask)))
 		return 0;
 
 	iter_info.srcu_idx = srcu_read_lock(&fsnotify_mark_srcu);
@@ -356,11 +366,13 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
 	if (mnt) {
 		iter_info.marks[FSNOTIFY_OBJ_TYPE_VFSMOUNT] =
 			fsnotify_first_mark(&mnt->mnt_fsnotify_marks);
+		iter_info.marks[FSNOTIFY_OBJ_TYPE_SB] =
+			fsnotify_first_mark(&sb->s_fsnotify_marks);
 	}
 
 	/*
-	 * We need to merge inode & vfsmount mark lists so that inode mark
-	 * ignore masks are properly reflected for mount mark notifications.
+	 * We need to merge inode/vfsmount/sb mark lists so that e.g. inode mark
+	 * ignore masks are properly reflected for mount/sb mark notifications.
 	 * That's why this traversal is so complicated...
 	 */
 	while (fsnotify_iter_select_report_types(&iter_info)) {
@@ -386,7 +398,7 @@ static __init int fsnotify_init(void)
 {
 	int ret;
 
-	BUG_ON(hweight32(ALL_FSNOTIFY_EVENTS) != 23);
+	BUILD_BUG_ON(HWEIGHT32(ALL_FSNOTIFY_BITS) != 23);
 
 	ret = init_srcu_struct(&fsnotify_mark_srcu);
 	if (ret)

fs/notify/fsnotify.h

@@ -21,6 +21,12 @@ static inline struct mount *fsnotify_conn_mount(
 	return container_of(conn->obj, struct mount, mnt_fsnotify_marks);
 }
 
+static inline struct super_block *fsnotify_conn_sb(
+				struct fsnotify_mark_connector *conn)
+{
+	return container_of(conn->obj, struct super_block, s_fsnotify_marks);
+}
+
 /* destroy all events sitting in this groups notification queue */
 extern void fsnotify_flush_notify(struct fsnotify_group *group);
 
@@ -43,6 +49,11 @@ static inline void fsnotify_clear_marks_by_mount(struct vfsmount *mnt)
 {
 	fsnotify_destroy_marks(&real_mount(mnt)->mnt_fsnotify_marks);
 }
+/* run the list of all marks associated with sb and destroy them */
+static inline void fsnotify_clear_marks_by_sb(struct super_block *sb)
+{
+	fsnotify_destroy_marks(&sb->s_fsnotify_marks);
+}
 /* Wait until all marks queued for destruction are destroyed */
 extern void fsnotify_wait_marks_destroyed(void);
 

fs/notify/inotify/inotify_user.c

@@ -815,7 +815,7 @@ static int __init inotify_user_setup(void)
 	BUILD_BUG_ON(IN_ISDIR != FS_ISDIR);
 	BUILD_BUG_ON(IN_ONESHOT != FS_IN_ONESHOT);
 
-	BUG_ON(hweight32(ALL_INOTIFY_BITS) != 22);
+	BUILD_BUG_ON(HWEIGHT32(ALL_INOTIFY_BITS) != 22);
 
 	inotify_inode_mark_cachep = KMEM_CACHE(inotify_inode_mark,
 					       SLAB_PANIC|SLAB_ACCOUNT);

fs/notify/mark.c

@@ -115,6 +115,8 @@ static __u32 *fsnotify_conn_mask_p(struct fsnotify_mark_connector *conn)
 		return &fsnotify_conn_inode(conn)->i_fsnotify_mask;
 	else if (conn->type == FSNOTIFY_OBJ_TYPE_VFSMOUNT)
 		return &fsnotify_conn_mount(conn)->mnt_fsnotify_mask;
+	else if (conn->type == FSNOTIFY_OBJ_TYPE_SB)
+		return &fsnotify_conn_sb(conn)->s_fsnotify_mask;
 	return NULL;
 }
 
@@ -179,19 +181,24 @@ static void fsnotify_connector_destroy_workfn(struct work_struct *work)
 	}
 }
 
-static struct inode *fsnotify_detach_connector_from_object(
-					struct fsnotify_mark_connector *conn)
+static void *fsnotify_detach_connector_from_object(
+					struct fsnotify_mark_connector *conn,
+					unsigned int *type)
 {
 	struct inode *inode = NULL;
 
+	*type = conn->type;
 	if (conn->type == FSNOTIFY_OBJ_TYPE_DETACHED)
 		return NULL;
 
 	if (conn->type == FSNOTIFY_OBJ_TYPE_INODE) {
 		inode = fsnotify_conn_inode(conn);
 		inode->i_fsnotify_mask = 0;
+		atomic_long_inc(&inode->i_sb->s_fsnotify_inode_refs);
 	} else if (conn->type == FSNOTIFY_OBJ_TYPE_VFSMOUNT) {
 		fsnotify_conn_mount(conn)->mnt_fsnotify_mask = 0;
+	} else if (conn->type == FSNOTIFY_OBJ_TYPE_SB) {
+		fsnotify_conn_sb(conn)->s_fsnotify_mask = 0;
 	}
 
 	rcu_assign_pointer(*(conn->obj), NULL);
@@ -211,10 +218,29 @@ static void fsnotify_final_mark_destroy(struct fsnotify_mark *mark)
 	fsnotify_put_group(group);
 }
 
+/* Drop object reference originally held by a connector */
+static void fsnotify_drop_object(unsigned int type, void *objp)
+{
+	struct inode *inode;
+	struct super_block *sb;
+
+	if (!objp)
+		return;
+	/* Currently only inode references are passed to be dropped */
+	if (WARN_ON_ONCE(type != FSNOTIFY_OBJ_TYPE_INODE))
+		return;
+	inode = objp;
+	sb = inode->i_sb;
+	iput(inode);
+	if (atomic_long_dec_and_test(&sb->s_fsnotify_inode_refs))
+		wake_up_var(&sb->s_fsnotify_inode_refs);
+}
+
 void fsnotify_put_mark(struct fsnotify_mark *mark)
 {
 	struct fsnotify_mark_connector *conn;
-	struct inode *inode = NULL;
+	void *objp = NULL;
+	unsigned int type = FSNOTIFY_OBJ_TYPE_DETACHED;
 	bool free_conn = false;
 
 	/* Catch marks that were actually never attached to object */
@@ -234,7 +260,7 @@ void fsnotify_put_mark(struct fsnotify_mark *mark)
 	conn = mark->connector;
 	hlist_del_init_rcu(&mark->obj_list);
 	if (hlist_empty(&conn->list)) {
-		inode = fsnotify_detach_connector_from_object(conn);
+		objp = fsnotify_detach_connector_from_object(conn, &type);
 		free_conn = true;
 	} else {
 		__fsnotify_recalc_mask(conn);
@@ -242,7 +268,7 @@ void fsnotify_put_mark(struct fsnotify_mark *mark)
 	mark->connector = NULL;
 	spin_unlock(&conn->lock);
 
-	iput(inode);
+	fsnotify_drop_object(type, objp);
 
 	if (free_conn) {
 		spin_lock(&destroy_lock);
@@ -709,7 +735,8 @@ void fsnotify_destroy_marks(fsnotify_connp_t *connp)
 {
 	struct fsnotify_mark_connector *conn;
 	struct fsnotify_mark *mark, *old_mark = NULL;
-	struct inode *inode;
+	void *objp;
+	unsigned int type;
 
 	conn = fsnotify_grab_connector(connp);
 	if (!conn)
@@ -735,11 +762,11 @@ void fsnotify_destroy_marks(fsnotify_connp_t *connp)
 	 * mark references get dropped. It would lead to strange results such
 	 * as delaying inode deletion or blocking unmount.
 	 */
-	inode = fsnotify_detach_connector_from_object(conn);
+	objp = fsnotify_detach_connector_from_object(conn, &type);
 	spin_unlock(&conn->lock);
 	if (old_mark)
 		fsnotify_put_mark(old_mark);
-	iput(inode);
+	fsnotify_drop_object(type, objp);
 }
 
 /*

fs/super.c

@@ -442,7 +442,7 @@ void generic_shutdown_super(struct super_block *sb)
 		sync_filesystem(sb);
 		sb->s_flags &= ~SB_ACTIVE;
 
-		fsnotify_unmount_inodes(sb);
+		fsnotify_sb_delete(sb);
 		cgroup_writeback_umount();
 
 		evict_inodes(sb);

include/linux/fanotify.h

@@ -4,6 +4,61 @@
 
 #include <uapi/linux/fanotify.h>
 
-/* not valid from userspace, only kernel internal */
-#define FAN_MARK_ONDIR		0x00000100
+#define FAN_GROUP_FLAG(group, flag) \
+	((group)->fanotify_data.flags & (flag))
+
+/*
+ * Flags allowed to be passed from/to userspace.
+ *
+ * We intentionally do not add new bits to the old FAN_ALL_* constants, because
+ * they are uapi exposed constants. If there are programs out there using
+ * these constant, the programs may break if re-compiled with new uapi headers
+ * and then run on an old kernel.
+ */
+#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_CLOEXEC | FAN_NONBLOCK | \
+				 FAN_UNLIMITED_QUEUE | FAN_UNLIMITED_MARKS)
+
+#define FANOTIFY_MARK_TYPE_BITS	(FAN_MARK_INODE | FAN_MARK_MOUNT | \
+				 FAN_MARK_FILESYSTEM)
+
+#define FANOTIFY_MARK_FLAGS	(FANOTIFY_MARK_TYPE_BITS | \
+				 FAN_MARK_ADD | \
+				 FAN_MARK_REMOVE | \
+				 FAN_MARK_DONT_FOLLOW | \
+				 FAN_MARK_ONLYDIR | \
+				 FAN_MARK_IGNORED_MASK | \
+				 FAN_MARK_IGNORED_SURV_MODIFY | \
+				 FAN_MARK_FLUSH)
+
+/* Events that user can request to be notified on */
+#define FANOTIFY_EVENTS		(FAN_ACCESS | FAN_MODIFY | \
+				 FAN_CLOSE | FAN_OPEN)
+
+/* Events that require a permission response from user */
+#define FANOTIFY_PERM_EVENTS	(FAN_OPEN_PERM | FAN_ACCESS_PERM)
+
+/* Extra flags that may be reported with event or control handling of events */
+#define FANOTIFY_EVENT_FLAGS	(FAN_EVENT_ON_CHILD | FAN_ONDIR)
+
+/* Events that may be reported to user */
+#define FANOTIFY_OUTGOING_EVENTS	(FANOTIFY_EVENTS | \
+					 FANOTIFY_PERM_EVENTS | \
+					 FAN_Q_OVERFLOW)
+
+#define ALL_FANOTIFY_EVENT_BITS		(FANOTIFY_OUTGOING_EVENTS | \
+					 FANOTIFY_EVENT_FLAGS)
+
+/* Do not use these old uapi constants internally */
+#undef FAN_ALL_CLASS_BITS
+#undef FAN_ALL_INIT_FLAGS
+#undef FAN_ALL_MARK_FLAGS
+#undef FAN_ALL_EVENTS
+#undef FAN_ALL_PERM_EVENTS
+#undef FAN_ALL_OUTGOING_EVENTS
+
 #endif /* _LINUX_FANOTIFY_H */

include/linux/fs.h

@@ -1412,17 +1412,26 @@ struct super_block {
 
 	struct sb_writers	s_writers;
 
+	/*
+	 * Keep s_fs_info, s_time_gran, s_fsnotify_mask, and
+	 * s_fsnotify_marks together for cache efficiency. They are frequently
+	 * accessed and rarely modified.
+	 */
+	void			*s_fs_info;	/* Filesystem private info */
+
+	/* Granularity of c/m/atime in ns (cannot be worse than a second) */
+	u32			s_time_gran;
+#ifdef CONFIG_FSNOTIFY
+	__u32			s_fsnotify_mask;
+	struct fsnotify_mark_connector __rcu	*s_fsnotify_marks;
+#endif
+
 	char			s_id[32];	/* Informational name */
 	uuid_t			s_uuid;		/* UUID */
 
-	void 			*s_fs_info;	/* Filesystem private info */
 	unsigned int		s_max_links;
 	fmode_t			s_mode;
 
-	/* Granularity of c/m/atime in ns.
-	   Cannot be worse than a second */
-	u32		   s_time_gran;
-
 	/*
 	 * The next field is for VFS *only*. No filesystems have any business
 	 * even looking at it. You had been warned.
@@ -1447,6 +1456,9 @@ struct super_block {
 	/* Number of inodes with nlink == 0 but still referenced */
 	atomic_long_t s_remove_count;
 
+	/* Pending fsnotify inode refs */
+	atomic_long_t s_fsnotify_inode_refs;
+
 	/* Being remounted read-only */
 	int s_readonly_remount;
 

include/linux/fsnotify_backend.h

@@ -68,15 +68,20 @@
 
 #define ALL_FSNOTIFY_PERM_EVENTS (FS_OPEN_PERM | FS_ACCESS_PERM)
 
+/* Events that can be reported to backends */
 #define ALL_FSNOTIFY_EVENTS (FS_ACCESS | FS_MODIFY | FS_ATTRIB | \
 			     FS_CLOSE_WRITE | FS_CLOSE_NOWRITE | FS_OPEN | \
 			     FS_MOVED_FROM | FS_MOVED_TO | FS_CREATE | \
 			     FS_DELETE | FS_DELETE_SELF | FS_MOVE_SELF | \
 			     FS_UNMOUNT | FS_Q_OVERFLOW | FS_IN_IGNORED | \
-			     FS_OPEN_PERM | FS_ACCESS_PERM | FS_EXCL_UNLINK | \
-			     FS_ISDIR | FS_IN_ONESHOT | FS_DN_RENAME | \
+			     FS_OPEN_PERM | FS_ACCESS_PERM | FS_DN_RENAME)
+
+/* Extra flags that may be reported with event or control handling of events */
+#define ALL_FSNOTIFY_FLAGS  (FS_EXCL_UNLINK | FS_ISDIR | FS_IN_ONESHOT | \
 			     FS_DN_MULTISHOT | FS_EVENT_ON_CHILD)
 
+#define ALL_FSNOTIFY_BITS   (ALL_FSNOTIFY_EVENTS | ALL_FSNOTIFY_FLAGS)
+
 struct fsnotify_group;
 struct fsnotify_event;
 struct fsnotify_mark;
@@ -189,10 +194,10 @@ struct fsnotify_group {
 			/* allows a group to block waiting for a userspace response */
 			struct list_head access_list;
 			wait_queue_head_t access_waitq;
-			int f_flags;
+			int flags;           /* flags from fanotify_init() */
+			int f_flags; /* event_f_flags from fanotify_init() */
 			unsigned int max_marks;
 			struct user_struct *user;
-			bool audit;
 		} fanotify_data;
 #endif /* CONFIG_FANOTIFY */
 	};
@@ -206,12 +211,14 @@ struct fsnotify_group {
 enum fsnotify_obj_type {
 	FSNOTIFY_OBJ_TYPE_INODE,
 	FSNOTIFY_OBJ_TYPE_VFSMOUNT,
+	FSNOTIFY_OBJ_TYPE_SB,
 	FSNOTIFY_OBJ_TYPE_COUNT,
 	FSNOTIFY_OBJ_TYPE_DETACHED = FSNOTIFY_OBJ_TYPE_COUNT
 };
 
 #define FSNOTIFY_OBJ_TYPE_INODE_FL	(1U << FSNOTIFY_OBJ_TYPE_INODE)
 #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)
 
 static inline bool fsnotify_valid_obj_type(unsigned int type)
@@ -255,6 +262,7 @@ static inline struct fsnotify_mark *fsnotify_iter_##name##_mark( \
 
 FSNOTIFY_ITER_FUNCS(inode, INODE)
 FSNOTIFY_ITER_FUNCS(vfsmount, VFSMOUNT)
+FSNOTIFY_ITER_FUNCS(sb, SB)
 
 #define fsnotify_foreach_obj_type(type) \
 	for (type = 0; type < FSNOTIFY_OBJ_TYPE_COUNT; type++)
@@ -267,8 +275,8 @@ struct fsnotify_mark_connector;
 typedef struct fsnotify_mark_connector __rcu *fsnotify_connp_t;
 
 /*
- * Inode / vfsmount point to this structure which tracks all marks attached to
- * the inode / vfsmount. The reference to inode / vfsmount is held by this
+ * Inode/vfsmount/sb point to this structure which tracks all marks attached to
+ * the inode/vfsmount/sb. The reference to inode/vfsmount/sb is held by this
  * structure. We destroy this structure when there are no more marks attached
  * to it. The structure is protected by fsnotify_mark_srcu.
  */
@@ -335,6 +343,7 @@ extern int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int dat
 extern int __fsnotify_parent(const struct path *path, struct dentry *dentry, __u32 mask);
 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 int fsnotify_inode_watches_children(struct inode *inode)
@@ -455,9 +464,13 @@ static inline void fsnotify_clear_inode_marks_by_group(struct fsnotify_group *gr
 {
 	fsnotify_clear_marks_by_group(group, FSNOTIFY_OBJ_TYPE_INODE_FL);
 }
+/* run all the marks in a group, and clear all of the sn marks */
+static inline void fsnotify_clear_sb_marks_by_group(struct fsnotify_group *group)
+{
+	fsnotify_clear_marks_by_group(group, FSNOTIFY_OBJ_TYPE_SB_FL);
+}
 extern void fsnotify_get_mark(struct fsnotify_mark *mark);
 extern void fsnotify_put_mark(struct fsnotify_mark *mark);
-extern void fsnotify_unmount_inodes(struct super_block *sb);
 extern void fsnotify_finish_user_wait(struct fsnotify_iter_info *iter_info);
 extern bool fsnotify_prepare_user_wait(struct fsnotify_iter_info *iter_info);
 
@@ -484,6 +497,9 @@ static inline void __fsnotify_inode_delete(struct inode *inode)
 static inline void __fsnotify_vfsmount_delete(struct vfsmount *mnt)
 {}
 
+static inline void fsnotify_sb_delete(struct super_block *sb)
+{}
+
 static inline void fsnotify_update_flags(struct dentry *dentry)
 {}
 

include/uapi/linux/fanotify.h

@@ -27,10 +27,12 @@
 #define FAN_CLOEXEC		0x00000001
 #define FAN_NONBLOCK		0x00000002
 
-/* These are NOT bitwise flags.  Both bits are used togther.  */
+/* These are NOT bitwise flags.  Both bits are used together.  */
 #define FAN_CLASS_NOTIF		0x00000000
 #define FAN_CLASS_CONTENT	0x00000004
 #define FAN_CLASS_PRE_CONTENT	0x00000008
+
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_CLASS_BITS	(FAN_CLASS_NOTIF | FAN_CLASS_CONTENT | \
 				 FAN_CLASS_PRE_CONTENT)
 
@@ -38,6 +40,10 @@
 #define FAN_UNLIMITED_MARKS	0x00000020
 #define FAN_ENABLE_AUDIT	0x00000040
 
+/* Flags to determine fanotify event format */
+#define FAN_REPORT_TID		0x00000100	/* event->pid is thread id */
+
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_INIT_FLAGS	(FAN_CLOEXEC | FAN_NONBLOCK | \
 				 FAN_ALL_CLASS_BITS | FAN_UNLIMITED_QUEUE |\
 				 FAN_UNLIMITED_MARKS)
@@ -47,11 +53,18 @@
 #define FAN_MARK_REMOVE		0x00000002
 #define FAN_MARK_DONT_FOLLOW	0x00000004
 #define FAN_MARK_ONLYDIR	0x00000008
-#define FAN_MARK_MOUNT		0x00000010
+/* FAN_MARK_MOUNT is		0x00000010 */
 #define FAN_MARK_IGNORED_MASK	0x00000020
 #define FAN_MARK_IGNORED_SURV_MODIFY	0x00000040
 #define FAN_MARK_FLUSH		0x00000080
+/* FAN_MARK_FILESYSTEM is	0x00000100 */
 
+/* These are NOT bitwise flags.  Both bits can be used togther.  */
+#define FAN_MARK_INODE		0x00000000
+#define FAN_MARK_MOUNT		0x00000010
+#define FAN_MARK_FILESYSTEM	0x00000100
+
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_MARK_FLAGS	(FAN_MARK_ADD |\
 				 FAN_MARK_REMOVE |\
 				 FAN_MARK_DONT_FOLLOW |\
@@ -61,11 +74,7 @@
 				 FAN_MARK_IGNORED_SURV_MODIFY |\
 				 FAN_MARK_FLUSH)
 
-/*
- * All of the events - we build the list by hand so that we can add flags in
- * the future and not break backward compatibility.  Apps will get only the
- * events that they originally wanted.  Be sure to add new events here!
- */
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_EVENTS (FAN_ACCESS |\
 			FAN_MODIFY |\
 			FAN_CLOSE |\
@@ -74,9 +83,11 @@
 /*
  * All events which require a permission response from userspace
  */
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_PERM_EVENTS (FAN_OPEN_PERM |\
 			     FAN_ACCESS_PERM)
 
+/* Deprecated - do not use this in programs and do not add new flags here! */
 #define FAN_ALL_OUTGOING_EVENTS	(FAN_ALL_EVENTS |\
 				 FAN_ALL_PERM_EVENTS |\
 				 FAN_Q_OVERFLOW)