fsnotify: parent event notification

inotify and dnotify both use a similar parent notification mechanism.  We
add a generic parent notification mechanism to fsnotify for both of these
to use.  This new machanism also adds the dentry flag optimization which
exists for inotify to dnotify.

Signed-off-by: Eric Paris <eparis@redhat.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>

3 files changed, 113 insertions, 0 deletions

diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c
index d5654629c659..7fc760067a62 100644
--- a/fs/notify/fsnotify.c
+++ b/fs/notify/fsnotify.c
@@ -35,6 +35,97 @@ void __fsnotify_inode_delete(struct inode *inode)
 EXPORT_SYMBOL_GPL(__fsnotify_inode_delete);
 
 /*
+ * Given an inode, first check if we care what happens to our children.  Inotify
+ * and dnotify both tell their parents about events.  If we care about any event
+ * on a child we run all of our children and set a dentry flag saying that the
+ * parent cares.  Thus when an event happens on a child it can quickly tell if
+ * if there is a need to find a parent and send the event to the parent.
+ */
+void __fsnotify_update_child_dentry_flags(struct inode *inode)
+{
+	struct dentry *alias;
+	int watched;
+
+	if (!S_ISDIR(inode->i_mode))
+		return;
+
+	/* determine if the children should tell inode about their events */
+	watched = fsnotify_inode_watches_children(inode);
+
+	spin_lock(&dcache_lock);
+	/* run all of the dentries associated with this inode.  Since this is a
+	 * directory, there damn well better only be one item on this list */
+	list_for_each_entry(alias, &inode->i_dentry, d_alias) {
+		struct dentry *child;
+
+		/* run all of the children of the original inode and fix their
+		 * d_flags to indicate parental interest (their parent is the
+		 * original inode) */
+		list_for_each_entry(child, &alias->d_subdirs, d_u.d_child) {
+			if (!child->d_inode)
+				continue;
+
+			spin_lock(&child->d_lock);
+			if (watched)
+				child->d_flags |= DCACHE_FSNOTIFY_PARENT_WATCHED;
+			else
+				child->d_flags &= ~DCACHE_FSNOTIFY_PARENT_WATCHED;
+			spin_unlock(&child->d_lock);
+		}
+	}
+	spin_unlock(&dcache_lock);
+}
+
+/* Notify this dentry's parent about a child's events. */
+void __fsnotify_parent(struct dentry *dentry, __u32 mask)
+{
+	struct dentry *parent;
+	struct inode *p_inode;
+	bool send = false;
+	bool should_update_children = false;
+
+	if (!(dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED))
+		return;
+
+	spin_lock(&dentry->d_lock);
+	parent = dentry->d_parent;
+	p_inode = parent->d_inode;
+
+	if (fsnotify_inode_watches_children(p_inode)) {
+		if (p_inode->i_fsnotify_mask & mask) {
+			dget(parent);
+			send = true;
+		}
+	} else {
+		/*
+		 * The parent doesn't care about events on it's children but
+		 * at least one child thought it did.  We need to run all the
+		 * children and update their d_flags to let them know p_inode
+		 * doesn't care about them any more.
+		 */
+		dget(parent);
+		should_update_children = true;
+	}
+
+	spin_unlock(&dentry->d_lock);
+
+	if (send) {
+		/* we are notifying a parent so come up with the new mask which
+		 * specifies these are events which came from a child. */
+		mask |= FS_EVENT_ON_CHILD;
+
+		fsnotify(p_inode, mask, dentry->d_inode, FSNOTIFY_EVENT_INODE);
+		dput(parent);
+	}
+
+	if (unlikely(should_update_children)) {
+		__fsnotify_update_child_dentry_flags(p_inode);
+		dput(parent);
+	}
+}
+EXPORT_SYMBOL_GPL(__fsnotify_parent);
+
+/*
  * This is the main call to fsnotify.  The VFS calls into hook specific functions
  * in linux/fsnotify.h.  Those functions then in turn call here.  Here will call
  * out to all of the registered fsnotify_group.  Those groups can then use the
diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h
index 8ebcbe893c91..83b8ec0a8ec2 100644
--- a/fs/notify/fsnotify.h
+++ b/fs/notify/fsnotify.h
@@ -17,4 +17,9 @@ extern __u32 fsnotify_mask;
 extern void fsnotify_final_destroy_group(struct fsnotify_group *group);
 /* run the list of all marks associated with inode and flag them to be freed */
 extern void fsnotify_clear_marks_by_inode(struct inode *inode);
+/*
+ * update the dentry->d_flags of all of inode's children to indicate if inode cares
+ * about events that happen to its children.
+ */
+extern void __fsnotify_update_child_dentry_flags(struct inode *inode);
 #endif	/* __FS_NOTIFY_FSNOTIFY_H_ */
diff --git a/fs/notify/inode_mark.c b/fs/notify/inode_mark.c
index cdc154146974..a39534845b28 100644
--- a/fs/notify/inode_mark.c
+++ b/fs/notify/inode_mark.c
@@ -131,6 +131,8 @@ void fsnotify_recalc_inode_mask(struct inode *inode)
 	spin_lock(&inode->i_lock);
 	fsnotify_recalc_inode_mask_locked(inode);
 	spin_unlock(&inode->i_lock);
+
+	__fsnotify_update_child_dentry_flags(inode);
 }
 
 /*
@@ -190,6 +192,19 @@ void fsnotify_destroy_mark_by_entry(struct fsnotify_mark_entry *entry)
 	group->ops->freeing_mark(entry, group);
 
 	/*
+	 * __fsnotify_update_child_dentry_flags(inode);
+	 *
+	 * I really want to call that, but we can't, we have no idea if the inode
+	 * still exists the second we drop the entry->lock.
+	 *
+	 * The next time an event arrive to this inode from one of it's children
+	 * __fsnotify_parent will see that the inode doesn't care about it's
+	 * children and will update all of these flags then.  So really this
+	 * is just a lazy update (and could be a perf win...)
+	 */
+
+
+	/*
 	 * it's possible that this group tried to destroy itself, but this
 	 * this mark was simultaneously being freed by inode.  If that's the
 	 * case, we finish freeing the group here.
@@ -323,6 +338,8 @@ int fsnotify_add_mark(struct fsnotify_mark_entry *entry,
 	if (lentry) {
 		ret = -EEXIST;
 		fsnotify_put_mark(lentry);
+	} else {
+		__fsnotify_update_child_dentry_flags(inode);
 	}
 
 	return ret;