fsnotify: send event to parent and child with single callback

Instead of calling fsnotify() twice, once with parent inode and once
with child inode, if event should be sent to parent inode, send it
with both parent and child inodes marks in object type iterator and call
the backend handle_event() callback only once.

The parent inode is assigned to the standard "inode" iterator type and
the child inode is assigned to the special "child" iterator type.

In that case, the bit FS_EVENT_ON_CHILD will be set in the event mask,
the dir argument to handle_event will be the parent inode, the file_name
argument to handle_event is non NULL and refers to the name of the child
and the child inode can be accessed with fsnotify_data_inode().

This will allow fanotify to make decisions based on child or parent's
ignored mask.  For example, when a parent is interested in a specific
event on its children, but a specific child wishes to ignore this event,
the event will not be reported.  This is not what happens with current
code, but according to man page, it is the expected behavior.

Link: https://lore.kernel.org/r/20200716084230.30611-15-amir73il@gmail.com
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>

3 files changed, 45 insertions, 33 deletions

diff --git a/fs/kernfs/file.c b/fs/kernfs/file.c
index e23b3f62483c..5b1468bc509e 100644
--- a/fs/kernfs/file.c
+++ b/fs/kernfs/file.c
@@ -883,6 +883,7 @@ repeat:
 
 	list_for_each_entry(info, &kernfs_root(kn)->supers, node) {
 		struct kernfs_node *parent;
+		struct inode *p_inode = NULL;
 		struct inode *inode;
 		struct qstr name;
 
@@ -899,8 +900,6 @@ repeat:
 		name = (struct qstr)QSTR_INIT(kn->name, strlen(kn->name));
 		parent = kernfs_get_parent(kn);
 		if (parent) {
-			struct inode *p_inode;
-
 			p_inode = ilookup(info->sb, kernfs_ino(parent));
 			if (p_inode) {
 				fsnotify(p_inode, FS_MODIFY | FS_EVENT_ON_CHILD,
@@ -911,8 +910,11 @@ repeat:
 			kernfs_put(parent);
 		}
 
-		fsnotify(inode, FS_MODIFY, inode, FSNOTIFY_EVENT_INODE,
-			 NULL, 0);
+		if (!p_inode) {
+			fsnotify(inode, FS_MODIFY, inode, FSNOTIFY_EVENT_INODE,
+				 NULL, 0);
+		}
+
 		iput(inode);
 	}
 
diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
index c107974d6830..3baf93e998c1 100644
--- a/fs/notify/fanotify/fanotify.c
+++ b/fs/notify/fanotify/fanotify.c
@@ -261,12 +261,12 @@ static u32 fanotify_group_event_mask(struct fsnotify_group *group,
 			continue;
 
 		/*
-		 * If the event is for a child and this mark doesn't care about
-		 * events on a child, don't send it!
+		 * If the event is for a child and this mark is on a parent not
+		 * watching children, don't send it!
 		 */
 		if (event_mask & FS_EVENT_ON_CHILD &&
-		    (type != FSNOTIFY_OBJ_TYPE_INODE ||
-		     !(mark->mask & FS_EVENT_ON_CHILD)))
+		    type == FSNOTIFY_OBJ_TYPE_INODE &&
+		     !(mark->mask & FS_EVENT_ON_CHILD))
 			continue;
 
 		marks_mask |= mark->mask;
diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c
index f12a554be3f0..0b0e01e04349 100644
--- a/fs/notify/fsnotify.c
+++ b/fs/notify/fsnotify.c
@@ -145,17 +145,21 @@ void __fsnotify_update_child_dentry_flags(struct inode *inode)
 /*
  * Notify this dentry's parent about a child's events with child name info
  * if parent is watching.
- * Notify also the child without name info if child inode is watching.
+ * Notify only the child without name info if parent is not watching.
  */
 int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
 		      int data_type)
 {
+	struct inode *inode = d_inode(dentry);
 	struct dentry *parent;
 	struct inode *p_inode;
+	struct name_snapshot name;
+	struct qstr *file_name = NULL;
 	int ret = 0;
 
+	parent = NULL;
 	if (!(dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED))
-		goto notify_child;
+		goto notify;
 
 	parent = dget_parent(dentry);
 	p_inode = parent->d_inode;
@@ -163,25 +167,24 @@ int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
 	if (unlikely(!fsnotify_inode_watches_children(p_inode))) {
 		__fsnotify_update_child_dentry_flags(p_inode);
 	} else if (p_inode->i_fsnotify_mask & mask & ALL_FSNOTIFY_EVENTS) {
-		struct name_snapshot name;
+		/* When notifying parent, child should be passed as data */
+		WARN_ON_ONCE(inode != fsnotify_data_inode(data, data_type));
 
-		/*
-		 * We are notifying a parent, so set a flag in mask to inform
-		 * backend that event has information about a child entry.
-		 */
+		/* Notify both parent and child with child name info */
+		inode = p_inode;
 		take_dentry_name_snapshot(&name, dentry);
-		ret = fsnotify(p_inode, mask | FS_EVENT_ON_CHILD, data,
-			       data_type, &name.name, 0);
-		release_dentry_name_snapshot(&name);
+		file_name = &name.name;
+		mask |= FS_EVENT_ON_CHILD;
 	}
 
-	dput(parent);
+notify:
+	ret = fsnotify(inode, mask, data, data_type, file_name, 0);
 
-	if (ret)
-		return ret;
+	if (file_name)
+		release_dentry_name_snapshot(&name);
+	dput(parent);
 
-notify_child:
-	return fsnotify(d_inode(dentry), mask, data, data_type, NULL, 0);
+	return ret;
 }
 EXPORT_SYMBOL_GPL(__fsnotify_parent);
 
@@ -322,12 +325,16 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_type,
 	struct super_block *sb = to_tell->i_sb;
 	struct inode *dir = file_name ? to_tell : NULL;
 	struct mount *mnt = NULL;
+	struct inode *child = NULL;
 	int ret = 0;
 	__u32 test_mask, marks_mask;
 
 	if (path)
 		mnt = real_mount(path->mnt);
 
+	if (mask & FS_EVENT_ON_CHILD)
+		child = fsnotify_data_inode(data, data_type);
+
 	/*
 	 * Optimization: srcu_read_lock() has a memory barrier which can
 	 * be expensive.  It protects walking the *_fsnotify_marks lists.
@@ -336,21 +343,20 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_type,
 	 * need SRCU to keep them "alive".
 	 */
 	if (!to_tell->i_fsnotify_marks && !sb->s_fsnotify_marks &&
-	    (!mnt || !mnt->mnt_fsnotify_marks))
+	    (!mnt || !mnt->mnt_fsnotify_marks) &&
+	    (!child || !child->i_fsnotify_marks))
 		return 0;
 
-	/* An event "on child" is not intended for a mount/sb mark */
-	marks_mask = to_tell->i_fsnotify_mask;
-	if (!(mask & FS_EVENT_ON_CHILD)) {
-		marks_mask |= sb->s_fsnotify_mask;
-		if (mnt)
-			marks_mask |= mnt->mnt_fsnotify_mask;
-	}
+	marks_mask = to_tell->i_fsnotify_mask | sb->s_fsnotify_mask;
+	if (mnt)
+		marks_mask |= mnt->mnt_fsnotify_mask;
+	if (child)
+		marks_mask |= child->i_fsnotify_mask;
+
 
 	/*
 	 * if this is a modify event we may need to clear the ignored masks
-	 * otherwise return if neither the inode nor the vfsmount/sb care about
-	 * this type of event.
+	 * otherwise return if none of the marks care about this type of event.
 	 */
 	test_mask = (mask & ALL_FSNOTIFY_EVENTS);
 	if (!(mask & FS_MODIFY) && !(test_mask & marks_mask))
@@ -366,6 +372,10 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_type,
 		iter_info.marks[FSNOTIFY_OBJ_TYPE_VFSMOUNT] =
 			fsnotify_first_mark(&mnt->mnt_fsnotify_marks);
 	}
+	if (child) {
+		iter_info.marks[FSNOTIFY_OBJ_TYPE_CHILD] =
+			fsnotify_first_mark(&child->i_fsnotify_marks);
+	}
 
 	/*
 	 * We need to merge inode/vfsmount/sb mark lists so that e.g. inode mark