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
  ...

fs/kernfs/file.c

@@ -883,6 +883,7 @@ static void kernfs_notify_workfn(struct work_struct *work)
 
 	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;
 
@@ -899,20 +900,20 @@ static void kernfs_notify_workfn(struct work_struct *work)
 		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,
-					 inode, FSNOTIFY_EVENT_INODE, &name, 0);
+				fsnotify(FS_MODIFY | FS_EVENT_ON_CHILD,
+					 inode, FSNOTIFY_EVENT_INODE,
+					 p_inode, &name, inode, 0);
 				iput(p_inode);
 			}
 
 			kernfs_put(parent);
 		}
 
-		fsnotify(inode, FS_MODIFY, inode, FSNOTIFY_EVENT_INODE,
-			 &name, 0);
+		if (!p_inode)
+			fsnotify_inode(inode, FS_MODIFY);
+
 		iput(inode);
 	}
 

fs/nfsd/filecache.c

@@ -598,11 +598,9 @@ static struct notifier_block nfsd_file_lease_notifier = {
 };
 
 static int
-nfsd_file_fsnotify_handle_event(struct fsnotify_group *group,
-				struct inode *inode,
-				u32 mask, const void *data, int data_type,
-				const struct qstr *file_name, u32 cookie,
-				struct fsnotify_iter_info *iter_info)
+nfsd_file_fsnotify_handle_event(struct fsnotify_mark *mark, u32 mask,
+				struct inode *inode, struct inode *dir,
+				const struct qstr *name)
 {
 	trace_nfsd_file_fsnotify_handle_event(inode, mask);
 
@@ -624,7 +622,7 @@ nfsd_file_fsnotify_handle_event(struct fsnotify_group *group,
 
 
 static const struct fsnotify_ops nfsd_file_fsnotify_ops = {
-	.handle_event = nfsd_file_fsnotify_handle_event,
+	.handle_inode_event = nfsd_file_fsnotify_handle_event,
 	.free_mark = nfsd_file_mark_free,
 };
 

fs/notify/dnotify/dnotify.c

@@ -70,13 +70,10 @@ static void dnotify_recalc_inode_mask(struct fsnotify_mark *fsn_mark)
  * destroy the dnotify struct if it was not registered to receive multiple
  * events.
  */
-static int dnotify_handle_event(struct fsnotify_group *group,
-				struct inode *inode,
-				u32 mask, const void *data, int data_type,
-				const struct qstr *file_name, u32 cookie,
-				struct fsnotify_iter_info *iter_info)
+static int dnotify_handle_event(struct fsnotify_mark *inode_mark, u32 mask,
+				struct inode *inode, struct inode *dir,
+				const struct qstr *name)
 {
-	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
 	struct dnotify_mark *dn_mark;
 	struct dnotify_struct *dn;
 	struct dnotify_struct **prev;
@@ -84,10 +81,7 @@ static int dnotify_handle_event(struct fsnotify_group *group,
 	__u32 test_mask = mask & ~FS_EVENT_ON_CHILD;
 
 	/* not a dir, dnotify doesn't care */
-	if (!S_ISDIR(inode->i_mode))
-		return 0;
-
-	if (WARN_ON(fsnotify_iter_vfsmount_mark(iter_info)))
+	if (!dir && !(mask & FS_ISDIR))
 		return 0;
 
 	dn_mark = container_of(inode_mark, struct dnotify_mark, fsn_mark);
@@ -127,7 +121,7 @@ static void dnotify_free_mark(struct fsnotify_mark *fsn_mark)
 }
 
 static const struct fsnotify_ops dnotify_fsnotify_ops = {
-	.handle_event = dnotify_handle_event,
+	.handle_inode_event = dnotify_handle_event,
 	.free_mark = dnotify_free_mark,
 };
 

fs/notify/fanotify/fanotify.c

@@ -34,10 +34,6 @@ static bool fanotify_fh_equal(struct fanotify_fh *fh1,
 	if (fh1->type != fh2->type || fh1->len != fh2->len)
 		return false;
 
-	/* Do not merge events if we failed to encode fh */
-	if (fh1->type == FILEID_INVALID)
-		return false;
-
 	return !fh1->len ||
 		!memcmp(fanotify_fh_buf(fh1), fanotify_fh_buf(fh2), fh1->len);
 }
@@ -53,25 +49,47 @@ static bool fanotify_fid_event_equal(struct fanotify_fid_event *ffe1,
 		fanotify_fh_equal(&ffe1->object_fh, &ffe2->object_fh);
 }
 
+static bool fanotify_info_equal(struct fanotify_info *info1,
+				struct fanotify_info *info2)
+{
+	if (info1->dir_fh_totlen != info2->dir_fh_totlen ||
+	    info1->file_fh_totlen != info2->file_fh_totlen ||
+	    info1->name_len != info2->name_len)
+		return false;
+
+	if (info1->dir_fh_totlen &&
+	    !fanotify_fh_equal(fanotify_info_dir_fh(info1),
+			       fanotify_info_dir_fh(info2)))
+		return false;
+
+	if (info1->file_fh_totlen &&
+	    !fanotify_fh_equal(fanotify_info_file_fh(info1),
+			       fanotify_info_file_fh(info2)))
+		return false;
+
+	return !info1->name_len ||
+		!memcmp(fanotify_info_name(info1), fanotify_info_name(info2),
+			info1->name_len);
+}
+
 static bool fanotify_name_event_equal(struct fanotify_name_event *fne1,
 				      struct fanotify_name_event *fne2)
 {
-	/*
-	 * Do not merge name events without dir fh.
-	 * FAN_DIR_MODIFY does not encode object fh, so it may be empty.
-	 */
-	if (!fne1->dir_fh.len)
+	struct fanotify_info *info1 = &fne1->info;
+	struct fanotify_info *info2 = &fne2->info;
+
+	/* Do not merge name events without dir fh */
+	if (!info1->dir_fh_totlen)
 		return false;
 
-	if (fne1->name_len != fne2->name_len ||
-	    !fanotify_fh_equal(&fne1->dir_fh, &fne2->dir_fh))
+	if (!fanotify_fsid_equal(&fne1->fsid, &fne2->fsid))
 		return false;
 
-	return !memcmp(fne1->name, fne2->name, fne1->name_len);
+	return fanotify_info_equal(info1, info2);
 }
 
 static bool fanotify_should_merge(struct fsnotify_event *old_fsn,
-			 struct fsnotify_event *new_fsn)
+				  struct fsnotify_event *new_fsn)
 {
 	struct fanotify_event *old, *new;
 
@@ -83,22 +101,22 @@ static bool fanotify_should_merge(struct fsnotify_event *old_fsn,
 	    old->type != new->type || old->pid != new->pid)
 		return false;
 
+	/*
+	 * We want to merge many dirent events in the same dir (i.e.
+	 * creates/unlinks/renames), but we do not want to merge dirent
+	 * events referring to subdirs with dirent events referring to
+	 * non subdirs, otherwise, user won't be able to tell from a
+	 * mask FAN_CREATE|FAN_DELETE|FAN_ONDIR if it describes mkdir+
+	 * unlink pair or rmdir+create pair of events.
+	 */
+	if ((old->mask & FS_ISDIR) != (new->mask & FS_ISDIR))
+		return false;
+
 	switch (old->type) {
 	case FANOTIFY_EVENT_TYPE_PATH:
 		return fanotify_path_equal(fanotify_event_path(old),
 					   fanotify_event_path(new));
 	case FANOTIFY_EVENT_TYPE_FID:
-		/*
-		 * We want to merge many dirent events in the same dir (i.e.
-		 * creates/unlinks/renames), but we do not want to merge dirent
-		 * events referring to subdirs with dirent events referring to
-		 * non subdirs, otherwise, user won't be able to tell from a
-		 * mask FAN_CREATE|FAN_DELETE|FAN_ONDIR if it describes mkdir+
-		 * unlink pair or rmdir+create pair of events.
-		 */
-		if ((old->mask & FS_ISDIR) != (new->mask & FS_ISDIR))
-			return false;
-
 		return fanotify_fid_event_equal(FANOTIFY_FE(old),
 						FANOTIFY_FE(new));
 	case FANOTIFY_EVENT_TYPE_FID_NAME:
@@ -208,24 +226,30 @@ static int fanotify_get_response(struct fsnotify_group *group,
 static u32 fanotify_group_event_mask(struct fsnotify_group *group,
 				     struct fsnotify_iter_info *iter_info,
 				     u32 event_mask, const void *data,
-				     int data_type)
+				     int data_type, struct inode *dir)
 {
 	__u32 marks_mask = 0, marks_ignored_mask = 0;
-	__u32 test_mask, user_mask = FANOTIFY_OUTGOING_EVENTS;
+	__u32 test_mask, user_mask = FANOTIFY_OUTGOING_EVENTS |
+				     FANOTIFY_EVENT_FLAGS;
 	const struct path *path = fsnotify_data_path(data, data_type);
+	unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
 	struct fsnotify_mark *mark;
 	int type;
 
 	pr_debug("%s: report_mask=%x mask=%x data=%p data_type=%d\n",
 		 __func__, iter_info->report_mask, event_mask, data, data_type);
 
-	if (!FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
+	if (!fid_mode) {
 		/* Do we have path to open a file descriptor? */
 		if (!path)
 			return 0;
 		/* Path type events are only relevant for files and dirs */
 		if (!d_is_reg(path->dentry) && !d_can_lookup(path->dentry))
 			return 0;
+	} else if (!(fid_mode & FAN_REPORT_FID)) {
+		/* Do we have a directory inode to report? */
+		if (!dir && !(event_mask & FS_ISDIR))
+			return 0;
 	}
 
 	fsnotify_foreach_obj_type(type) {
@@ -244,12 +268,12 @@ 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;
@@ -264,67 +288,102 @@ static u32 fanotify_group_event_mask(struct fsnotify_group *group,
 	 *
 	 * For backward compatibility and consistency, do not report FAN_ONDIR
 	 * to user in legacy fanotify mode (reporting fd) and report FAN_ONDIR
-	 * to user in FAN_REPORT_FID mode for all event types.
+	 * to user in fid mode for all event types.
+	 *
+	 * We never report FAN_EVENT_ON_CHILD to user, but we do pass it in to
+	 * fanotify_alloc_event() when group is reporting fid as indication
+	 * that event happened on child.
 	 */
-	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
-		/* Do not report FAN_ONDIR without any event */
-		if (!(test_mask & ~FAN_ONDIR))
+	if (fid_mode) {
+		/* Do not report event flags without any event */
+		if (!(test_mask & ~FANOTIFY_EVENT_FLAGS))
 			return 0;
 	} else {
-		user_mask &= ~FAN_ONDIR;
+		user_mask &= ~FANOTIFY_EVENT_FLAGS;
 	}
 
 	return test_mask & user_mask;
 }
 
-static void fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
-			       gfp_t gfp)
+/*
+ * Check size needed to encode fanotify_fh.
+ *
+ * Return size of encoded fh without fanotify_fh header.
+ * Return 0 on failure to encode.
+ */
+static int fanotify_encode_fh_len(struct inode *inode)
 {
-	int dwords, type, bytes = 0;
+	int dwords = 0;
+
+	if (!inode)
+		return 0;
+
+	exportfs_encode_inode_fh(inode, NULL, &dwords, NULL);
+
+	return dwords << 2;
+}
+
+/*
+ * Encode fanotify_fh.
+ *
+ * Return total size of encoded fh including fanotify_fh header.
+ * Return 0 on failure to encode.
+ */
+static int fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
+			      unsigned int fh_len, gfp_t gfp)
+{
+	int dwords, type = 0;
 	char *ext_buf = NULL;
 	void *buf = fh->buf;
 	int err;
 
+	fh->type = FILEID_ROOT;
+	fh->len = 0;
+	fh->flags = 0;
 	if (!inode)
-		goto out;
+		return 0;
 
-	dwords = 0;
+	/*
+	 * !gpf means preallocated variable size fh, but fh_len could
+	 * be zero in that case if encoding fh len failed.
+	 */
 	err = -ENOENT;
-	type = exportfs_encode_inode_fh(inode, NULL, &dwords, NULL);
-	if (!dwords)
+	if (fh_len < 4 || WARN_ON_ONCE(fh_len % 4))
 		goto out_err;
 
-	bytes = dwords << 2;
-	if (bytes > FANOTIFY_INLINE_FH_LEN) {
-		/* Treat failure to allocate fh as failure to allocate event */
+	/* No external buffer in a variable size allocated fh */
+	if (gfp && fh_len > FANOTIFY_INLINE_FH_LEN) {
+		/* Treat failure to allocate fh as failure to encode fh */
 		err = -ENOMEM;
-		ext_buf = kmalloc(bytes, gfp);
+		ext_buf = kmalloc(fh_len, gfp);
 		if (!ext_buf)
 			goto out_err;
 
 		*fanotify_fh_ext_buf_ptr(fh) = ext_buf;
 		buf = ext_buf;
+		fh->flags |= FANOTIFY_FH_FLAG_EXT_BUF;
 	}
 
+	dwords = fh_len >> 2;
 	type = exportfs_encode_inode_fh(inode, buf, &dwords, NULL);
 	err = -EINVAL;
-	if (!type || type == FILEID_INVALID || bytes != dwords << 2)
+	if (!type || type == FILEID_INVALID || fh_len != dwords << 2)
 		goto out_err;
 
 	fh->type = type;
-	fh->len = bytes;
+	fh->len = fh_len;
 
-	return;
+	return FANOTIFY_FH_HDR_LEN + fh_len;
 
 out_err:
 	pr_warn_ratelimited("fanotify: failed to encode fid (type=%d, len=%d, err=%i)\n",
-			    type, bytes, err);
+			    type, fh_len, err);
 	kfree(ext_buf);
 	*fanotify_fh_ext_buf_ptr(fh) = NULL;
-out:
 	/* Report the event without a file identifier on encode error */
 	fh->type = FILEID_INVALID;
 	fh->len = 0;
+	return 0;
 }
 
 /*
@@ -335,27 +394,179 @@ static void fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
  * FS_ATTRIB reports the child inode even if reported on a watched parent.
  * FS_CREATE reports the modified dir inode and not the created inode.
  */
-static struct inode *fanotify_fid_inode(struct inode *to_tell, u32 event_mask,
-					const void *data, int data_type)
+static struct inode *fanotify_fid_inode(u32 event_mask, const void *data,
+					int data_type, struct inode *dir)
 {
 	if (event_mask & ALL_FSNOTIFY_DIRENT_EVENTS)
-		return to_tell;
+		return dir;
+
+	return fsnotify_data_inode(data, data_type);
+}
+
+/*
+ * The inode to use as identifier when reporting dir fid depends on the event.
+ * Report the modified directory inode on dirent modification events.
+ * Report the "victim" inode if "victim" is a directory.
+ * Report the parent inode if "victim" is not a directory and event is
+ * reported to parent.
+ * Otherwise, do not report dir fid.
+ */
+static struct inode *fanotify_dfid_inode(u32 event_mask, const void *data,
+					 int data_type, struct inode *dir)
+{
+	struct inode *inode = fsnotify_data_inode(data, data_type);
+
+	if (event_mask & ALL_FSNOTIFY_DIRENT_EVENTS)
+		return dir;
+
+	if (S_ISDIR(inode->i_mode))
+		return inode;
+
+	return dir;
+}
+
+static struct fanotify_event *fanotify_alloc_path_event(const struct path *path,
+							gfp_t gfp)
+{
+	struct fanotify_path_event *pevent;
+
+	pevent = kmem_cache_alloc(fanotify_path_event_cachep, gfp);
+	if (!pevent)
+		return NULL;
+
+	pevent->fae.type = FANOTIFY_EVENT_TYPE_PATH;
+	pevent->path = *path;
+	path_get(path);
+
+	return &pevent->fae;
+}
 
-	return (struct inode *)fsnotify_data_inode(data, data_type);
+static struct fanotify_event *fanotify_alloc_perm_event(const struct path *path,
+							gfp_t gfp)
+{
+	struct fanotify_perm_event *pevent;
+
+	pevent = kmem_cache_alloc(fanotify_perm_event_cachep, gfp);
+	if (!pevent)
+		return NULL;
+
+	pevent->fae.type = FANOTIFY_EVENT_TYPE_PATH_PERM;
+	pevent->response = 0;
+	pevent->state = FAN_EVENT_INIT;
+	pevent->path = *path;
+	path_get(path);
+
+	return &pevent->fae;
 }
 
-struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
-					    struct inode *inode, u32 mask,
-					    const void *data, int data_type,
-					    const struct qstr *file_name,
-					    __kernel_fsid_t *fsid)
+static struct fanotify_event *fanotify_alloc_fid_event(struct inode *id,
+						       __kernel_fsid_t *fsid,
+						       gfp_t gfp)
+{
+	struct fanotify_fid_event *ffe;
+
+	ffe = kmem_cache_alloc(fanotify_fid_event_cachep, gfp);
+	if (!ffe)
+		return NULL;
+
+	ffe->fae.type = FANOTIFY_EVENT_TYPE_FID;
+	ffe->fsid = *fsid;
+	fanotify_encode_fh(&ffe->object_fh, id, fanotify_encode_fh_len(id),
+			   gfp);
+
+	return &ffe->fae;
+}
+
+static struct fanotify_event *fanotify_alloc_name_event(struct inode *id,
+							__kernel_fsid_t *fsid,
+							const struct qstr *file_name,
+							struct inode *child,
+							gfp_t gfp)
+{
+	struct fanotify_name_event *fne;
+	struct fanotify_info *info;
+	struct fanotify_fh *dfh, *ffh;
+	unsigned int dir_fh_len = fanotify_encode_fh_len(id);
+	unsigned int child_fh_len = fanotify_encode_fh_len(child);
+	unsigned int size;
+
+	size = sizeof(*fne) + FANOTIFY_FH_HDR_LEN + dir_fh_len;
+	if (child_fh_len)
+		size += FANOTIFY_FH_HDR_LEN + child_fh_len;
+	if (file_name)
+		size += file_name->len + 1;
+	fne = kmalloc(size, gfp);
+	if (!fne)
+		return NULL;
+
+	fne->fae.type = FANOTIFY_EVENT_TYPE_FID_NAME;
+	fne->fsid = *fsid;
+	info = &fne->info;
+	fanotify_info_init(info);
+	dfh = fanotify_info_dir_fh(info);
+	info->dir_fh_totlen = fanotify_encode_fh(dfh, id, dir_fh_len, 0);
+	if (child_fh_len) {
+		ffh = fanotify_info_file_fh(info);
+		info->file_fh_totlen = fanotify_encode_fh(ffh, child, child_fh_len, 0);
+	}
+	if (file_name)
+		fanotify_info_copy_name(info, file_name);
+
+	pr_debug("%s: ino=%lu size=%u dir_fh_len=%u child_fh_len=%u name_len=%u name='%.*s'\n",
+		 __func__, id->i_ino, size, dir_fh_len, child_fh_len,
+		 info->name_len, info->name_len, fanotify_info_name(info));
+
+	return &fne->fae;
+}
+
+static struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
+						   u32 mask, const void *data,
+						   int data_type, struct inode *dir,
+						   const struct qstr *file_name,
+						   __kernel_fsid_t *fsid)
 {
 	struct fanotify_event *event = NULL;
-	struct fanotify_fid_event *ffe = NULL;
-	struct fanotify_name_event *fne = NULL;
 	gfp_t gfp = GFP_KERNEL_ACCOUNT;
-	struct inode *id = fanotify_fid_inode(inode, mask, data, data_type);
+	struct inode *id = fanotify_fid_inode(mask, data, data_type, dir);
+	struct inode *dirid = fanotify_dfid_inode(mask, data, data_type, dir);
 	const struct path *path = fsnotify_data_path(data, data_type);
+	unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
+	struct inode *child = NULL;
+	bool name_event = false;
+
+	if ((fid_mode & FAN_REPORT_DIR_FID) && dirid) {
+		/*
+		 * With both flags FAN_REPORT_DIR_FID and FAN_REPORT_FID, we
+		 * report the child fid for events reported on a non-dir child
+		 * in addition to reporting the parent fid and maybe child name.
+		 */
+		if ((fid_mode & FAN_REPORT_FID) &&
+		    id != dirid && !(mask & FAN_ONDIR))
+			child = id;
+
+		id = dirid;
+
+		/*
+		 * We record file name only in a group with FAN_REPORT_NAME
+		 * and when we have a directory inode to report.
+		 *
+		 * For directory entry modification event, we record the fid of
+		 * the directory and the name of the modified entry.
+		 *
+		 * For event on non-directory that is reported to parent, we
+		 * record the fid of the parent and the name of the child.
+		 *
+		 * Even if not reporting name, we need a variable length
+		 * fanotify_name_event if reporting both parent and child fids.
+		 */
+		if (!(fid_mode & FAN_REPORT_NAME)) {
+			name_event = !!child;
+			file_name = NULL;
+		} else if ((mask & ALL_FSNOTIFY_DIRENT_EVENTS) ||
+			   !(mask & FAN_ONDIR)) {
+			name_event = true;
+		}
+	}
 
 	/*
 	 * For queues with unlimited length lost events are not expected and
@@ -372,87 +583,30 @@ struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
 	memalloc_use_memcg(group->memcg);
 
 	if (fanotify_is_perm_event(mask)) {
-		struct fanotify_perm_event *pevent;
-
-		pevent = kmem_cache_alloc(fanotify_perm_event_cachep, gfp);
-		if (!pevent)
-			goto out;
-
-		event = &pevent->fae;
-		event->type = FANOTIFY_EVENT_TYPE_PATH_PERM;
-		pevent->response = 0;
-		pevent->state = FAN_EVENT_INIT;
-		goto init;
-	}
-
-	/*
-	 * For FAN_DIR_MODIFY event, we report the fid of the directory and
-	 * the name of the modified entry.
-	 * Allocate an fanotify_name_event struct and copy the name.
-	 */
-	if (mask & FAN_DIR_MODIFY && !(WARN_ON_ONCE(!file_name))) {
-		fne = kmalloc(sizeof(*fne) + file_name->len + 1, gfp);
-		if (!fne)
-			goto out;
-
-		event = &fne->fae;
-		event->type = FANOTIFY_EVENT_TYPE_FID_NAME;
-		fne->name_len = file_name->len;
-		strcpy(fne->name, file_name->name);
-		goto init;
-	}
-
-	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
-		ffe = kmem_cache_alloc(fanotify_fid_event_cachep, gfp);
-		if (!ffe)
-			goto out;
-
-		event = &ffe->fae;
-		event->type = FANOTIFY_EVENT_TYPE_FID;
+		event = fanotify_alloc_perm_event(path, gfp);
+	} else if (name_event && (file_name || child)) {
+		event = fanotify_alloc_name_event(id, fsid, file_name, child,
+						  gfp);
+	} else if (fid_mode) {
+		event = fanotify_alloc_fid_event(id, fsid, gfp);
 	} else {
-		struct fanotify_path_event *pevent;
-
-		pevent = kmem_cache_alloc(fanotify_path_event_cachep, gfp);
-		if (!pevent)
-			goto out;
-
-		event = &pevent->fae;
-		event->type = FANOTIFY_EVENT_TYPE_PATH;
+		event = fanotify_alloc_path_event(path, gfp);
 	}
 
-init:
+	if (!event)
+		goto out;
+
 	/*
 	 * Use the victim inode instead of the watching inode as the id for
 	 * event queue, so event reported on parent is merged with event
 	 * reported on child when both directory and child watches exist.
 	 */
-	fsnotify_init_event(&event->fse, (unsigned long)id);
-	event->mask = mask;
+	fanotify_init_event(event, (unsigned long)id, mask);
 	if (FAN_GROUP_FLAG(group, FAN_REPORT_TID))
 		event->pid = get_pid(task_pid(current));
 	else
 		event->pid = get_pid(task_tgid(current));
 
-	if (fsid && fanotify_event_fsid(event))
-		*fanotify_event_fsid(event) = *fsid;
-
-	if (fanotify_event_object_fh(event))
-		fanotify_encode_fh(fanotify_event_object_fh(event), id, gfp);
-
-	if (fanotify_event_dir_fh(event))
-		fanotify_encode_fh(fanotify_event_dir_fh(event), id, gfp);
-
-	if (fanotify_event_has_path(event)) {
-		struct path *p = fanotify_event_path(event);
-
-		if (path) {
-			*p = *path;
-			path_get(path);
-		} else {
-			p->mnt = NULL;
-			p->dentry = NULL;
-		}
-	}
 out:
 	memalloc_unuse_memcg();
 	return event;
@@ -491,9 +645,9 @@ static __kernel_fsid_t fanotify_get_fsid(struct fsnotify_iter_info *iter_info)
 	return fsid;
 }
 
-static int fanotify_handle_event(struct fsnotify_group *group,
-				 struct inode *inode,
-				 u32 mask, const void *data, int data_type,
+static int fanotify_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)
 {
@@ -512,7 +666,6 @@ static int fanotify_handle_event(struct fsnotify_group *group,
 	BUILD_BUG_ON(FAN_MOVED_FROM != FS_MOVED_FROM);
 	BUILD_BUG_ON(FAN_CREATE != FS_CREATE);
 	BUILD_BUG_ON(FAN_DELETE != FS_DELETE);
-	BUILD_BUG_ON(FAN_DIR_MODIFY != FS_DIR_MODIFY);
 	BUILD_BUG_ON(FAN_DELETE_SELF != FS_DELETE_SELF);
 	BUILD_BUG_ON(FAN_MOVE_SELF != FS_MOVE_SELF);
 	BUILD_BUG_ON(FAN_EVENT_ON_CHILD != FS_EVENT_ON_CHILD);
@@ -526,12 +679,11 @@ static int fanotify_handle_event(struct fsnotify_group *group,
 	BUILD_BUG_ON(HWEIGHT32(ALL_FANOTIFY_EVENT_BITS) != 19);
 
 	mask = fanotify_group_event_mask(group, iter_info, mask, data,
-					 data_type);
+					 data_type, dir);
 	if (!mask)
 		return 0;
 
-	pr_debug("%s: group=%p inode=%p mask=%x\n", __func__, group, inode,
-		 mask);
+	pr_debug("%s: group=%p mask=%x\n", __func__, group, mask);
 
 	if (fanotify_is_perm_event(mask)) {
 		/*
@@ -542,14 +694,14 @@ static int fanotify_handle_event(struct fsnotify_group *group,
 			return 0;
 	}
 
-	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
+	if (FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS)) {
 		fsid = fanotify_get_fsid(iter_info);
 		/* Racing with mark destruction or creation? */
 		if (!fsid.val[0] && !fsid.val[1])
 			return 0;
 	}
 
-	event = fanotify_alloc_event(group, inode, mask, data, data_type,
+	event = fanotify_alloc_event(group, mask, data, data_type, dir,
 				     file_name, &fsid);
 	ret = -ENOMEM;
 	if (unlikely(!event)) {
@@ -614,11 +766,7 @@ static void fanotify_free_fid_event(struct fanotify_event *event)
 
 static void fanotify_free_name_event(struct fanotify_event *event)
 {
-	struct fanotify_name_event *fne = FANOTIFY_NE(event);
-
-	if (fanotify_fh_has_ext_buf(&fne->dir_fh))
-		kfree(fanotify_fh_ext_buf(&fne->dir_fh));
-	kfree(fne);
+	kfree(FANOTIFY_NE(event));
 }
 
 static void fanotify_free_event(struct fsnotify_event *fsn_event)
@@ -640,6 +788,9 @@ static void fanotify_free_event(struct fsnotify_event *fsn_event)
 	case FANOTIFY_EVENT_TYPE_FID_NAME:
 		fanotify_free_name_event(event);
 		break;
+	case FANOTIFY_EVENT_TYPE_OVERFLOW:
+		kfree(event);
+		break;
 	default:
 		WARN_ON_ONCE(1);
 	}

fs/notify/fanotify/fanotify.h

@@ -23,20 +23,41 @@ enum {
  * stored in either the first or last 2 dwords.
  */
 #define FANOTIFY_INLINE_FH_LEN	(3 << 2)
+#define FANOTIFY_FH_HDR_LEN	offsetof(struct fanotify_fh, buf)
 
+/* Fixed size struct for file handle */
 struct fanotify_fh {
-	unsigned char buf[FANOTIFY_INLINE_FH_LEN];
 	u8 type;
 	u8 len;
+#define FANOTIFY_FH_FLAG_EXT_BUF 1
+	u8 flags;
+	u8 pad;
+	unsigned char buf[];
+} __aligned(4);
+
+/* Variable size struct for dir file handle + child file handle + name */
+struct fanotify_info {
+	/* size of dir_fh/file_fh including fanotify_fh hdr size */
+	u8 dir_fh_totlen;
+	u8 file_fh_totlen;
+	u8 name_len;
+	u8 pad;
+	unsigned char buf[];
+	/*
+	 * (struct fanotify_fh) dir_fh starts at buf[0]
+	 * (optional) file_fh starts at buf[dir_fh_totlen]
+	 * name starts at buf[dir_fh_totlen + file_fh_totlen]
+	 */
 } __aligned(4);
 
 static inline bool fanotify_fh_has_ext_buf(struct fanotify_fh *fh)
 {
-	return fh->len > FANOTIFY_INLINE_FH_LEN;
+	return (fh->flags & FANOTIFY_FH_FLAG_EXT_BUF);
 }
 
 static inline char **fanotify_fh_ext_buf_ptr(struct fanotify_fh *fh)
 {
+	BUILD_BUG_ON(FANOTIFY_FH_HDR_LEN % 4);
 	BUILD_BUG_ON(__alignof__(char *) - 4 + sizeof(char *) >
 		     FANOTIFY_INLINE_FH_LEN);
 	return (char **)ALIGN((unsigned long)(fh->buf), __alignof__(char *));
@@ -52,6 +73,56 @@ static inline void *fanotify_fh_buf(struct fanotify_fh *fh)
 	return fanotify_fh_has_ext_buf(fh) ? fanotify_fh_ext_buf(fh) : fh->buf;
 }
 
+static inline int fanotify_info_dir_fh_len(struct fanotify_info *info)
+{
+	if (!info->dir_fh_totlen ||
+	    WARN_ON_ONCE(info->dir_fh_totlen < FANOTIFY_FH_HDR_LEN))
+		return 0;
+
+	return info->dir_fh_totlen - FANOTIFY_FH_HDR_LEN;
+}
+
+static inline struct fanotify_fh *fanotify_info_dir_fh(struct fanotify_info *info)
+{
+	BUILD_BUG_ON(offsetof(struct fanotify_info, buf) % 4);
+
+	return (struct fanotify_fh *)info->buf;
+}
+
+static inline int fanotify_info_file_fh_len(struct fanotify_info *info)
+{
+	if (!info->file_fh_totlen ||
+	    WARN_ON_ONCE(info->file_fh_totlen < FANOTIFY_FH_HDR_LEN))
+		return 0;
+
+	return info->file_fh_totlen - FANOTIFY_FH_HDR_LEN;
+}
+
+static inline struct fanotify_fh *fanotify_info_file_fh(struct fanotify_info *info)
+{
+	return (struct fanotify_fh *)(info->buf + info->dir_fh_totlen);
+}
+
+static inline const char *fanotify_info_name(struct fanotify_info *info)
+{
+	return info->buf + info->dir_fh_totlen + info->file_fh_totlen;
+}
+
+static inline void fanotify_info_init(struct fanotify_info *info)
+{
+	info->dir_fh_totlen = 0;
+	info->file_fh_totlen = 0;
+	info->name_len = 0;
+}
+
+static inline void fanotify_info_copy_name(struct fanotify_info *info,
+					   const struct qstr *name)
+{
+	info->name_len = name->len;
+	strcpy(info->buf + info->dir_fh_totlen + info->file_fh_totlen,
+	       name->name);
+}
+
 /*
  * Common structure for fanotify events. Concrete structs are allocated in
  * fanotify_handle_event() and freed when the information is retrieved by
@@ -63,6 +134,7 @@ enum fanotify_event_type {
 	FANOTIFY_EVENT_TYPE_FID_NAME, /* variable length */
 	FANOTIFY_EVENT_TYPE_PATH,
 	FANOTIFY_EVENT_TYPE_PATH_PERM,
+	FANOTIFY_EVENT_TYPE_OVERFLOW, /* struct fanotify_event */
 };
 
 struct fanotify_event {
@@ -72,10 +144,20 @@ struct fanotify_event {
 	struct pid *pid;
 };
 
+static inline void fanotify_init_event(struct fanotify_event *event,
+				       unsigned long id, u32 mask)
+{
+	fsnotify_init_event(&event->fse, id);
+	event->mask = mask;
+	event->pid = NULL;
+}
+
 struct fanotify_fid_event {
 	struct fanotify_event fae;
 	__kernel_fsid_t fsid;
 	struct fanotify_fh object_fh;
+	/* Reserve space in object_fh.buf[] - access with fanotify_fh_buf() */
+	unsigned char _inline_fh_buf[FANOTIFY_INLINE_FH_LEN];
 };
 
 static inline struct fanotify_fid_event *
@@ -87,9 +169,7 @@ FANOTIFY_FE(struct fanotify_event *event)
 struct fanotify_name_event {
 	struct fanotify_event fae;
 	__kernel_fsid_t fsid;
-	struct fanotify_fh dir_fh;
-	u8 name_len;
-	char name[];
+	struct fanotify_info info;
 };
 
 static inline struct fanotify_name_event *
@@ -113,35 +193,37 @@ static inline struct fanotify_fh *fanotify_event_object_fh(
 {
 	if (event->type == FANOTIFY_EVENT_TYPE_FID)
 		return &FANOTIFY_FE(event)->object_fh;
+	else if (event->type == FANOTIFY_EVENT_TYPE_FID_NAME)
+		return fanotify_info_file_fh(&FANOTIFY_NE(event)->info);
 	else
 		return NULL;
 }
 
-static inline struct fanotify_fh *fanotify_event_dir_fh(
+static inline struct fanotify_info *fanotify_event_info(
 						struct fanotify_event *event)
 {
 	if (event->type == FANOTIFY_EVENT_TYPE_FID_NAME)
-		return &FANOTIFY_NE(event)->dir_fh;
+		return &FANOTIFY_NE(event)->info;
 	else
 		return NULL;
 }
 
 static inline int fanotify_event_object_fh_len(struct fanotify_event *event)
 {
+	struct fanotify_info *info = fanotify_event_info(event);
 	struct fanotify_fh *fh = fanotify_event_object_fh(event);
 
-	return fh ? fh->len : 0;
+	if (info)
+		return info->file_fh_totlen ? fh->len : 0;
+	else
+		return fh ? fh->len : 0;
 }
 
-static inline bool fanotify_event_has_name(struct fanotify_event *event)
+static inline int fanotify_event_dir_fh_len(struct fanotify_event *event)
 {
-	return event->type == FANOTIFY_EVENT_TYPE_FID_NAME;
-}
+	struct fanotify_info *info = fanotify_event_info(event);
 
-static inline int fanotify_event_name_len(struct fanotify_event *event)
-{
-	return fanotify_event_has_name(event) ?
-		FANOTIFY_NE(event)->name_len : 0;
+	return info ? fanotify_info_dir_fh_len(info) : 0;
 }
 
 struct fanotify_path_event {
@@ -202,9 +284,3 @@ static inline struct path *fanotify_event_path(struct fanotify_event *event)
 	else
 		return NULL;
 }
-
-struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
-					    struct inode *inode, u32 mask,
-					    const void *data, int data_type,
-					    const struct qstr *file_name,
-					    __kernel_fsid_t *fsid);

fs/notify/fanotify/fanotify_user.c

@@ -64,21 +64,28 @@ static int fanotify_fid_info_len(int fh_len, int name_len)
 	return roundup(FANOTIFY_INFO_HDR_LEN + info_len, FANOTIFY_EVENT_ALIGN);
 }
 
-static int fanotify_event_info_len(struct fanotify_event *event)
+static int fanotify_event_info_len(unsigned int fid_mode,
+				   struct fanotify_event *event)
 {
-	int info_len = 0;
+	struct fanotify_info *info = fanotify_event_info(event);
+	int dir_fh_len = fanotify_event_dir_fh_len(event);
 	int fh_len = fanotify_event_object_fh_len(event);
+	int info_len = 0;
+	int dot_len = 0;
 
-	if (fh_len)
-		info_len += fanotify_fid_info_len(fh_len, 0);
-
-	if (fanotify_event_name_len(event)) {
-		struct fanotify_name_event *fne = FANOTIFY_NE(event);
-
-		info_len += fanotify_fid_info_len(fne->dir_fh.len,
-						  fne->name_len);
+	if (dir_fh_len) {
+		info_len += fanotify_fid_info_len(dir_fh_len, info->name_len);
+	} else if ((fid_mode & FAN_REPORT_NAME) && (event->mask & FAN_ONDIR)) {
+		/*
+		 * With group flag FAN_REPORT_NAME, if name was not recorded in
+		 * event on a directory, we will report the name ".".
+		 */
+		dot_len = 1;
 	}
 
+	if (fh_len)
+		info_len += fanotify_fid_info_len(fh_len, dot_len);
+
 	return info_len;
 }
 
@@ -93,6 +100,7 @@ static struct fanotify_event *get_one_event(struct fsnotify_group *group,
 {
 	size_t event_size = FAN_EVENT_METADATA_LEN;
 	struct fanotify_event *event = NULL;
+	unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
 
 	pr_debug("%s: group=%p count=%zd\n", __func__, group, count);
 
@@ -100,8 +108,8 @@ static struct fanotify_event *get_one_event(struct fsnotify_group *group,
 	if (fsnotify_notify_queue_is_empty(group))
 		goto out;
 
-	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
-		event_size += fanotify_event_info_len(
+	if (fid_mode) {
+		event_size += fanotify_event_info_len(fid_mode,
 			FANOTIFY_E(fsnotify_peek_first_event(group)));
 	}
 
@@ -218,7 +226,7 @@ static int process_access_response(struct fsnotify_group *group,
 }
 
 static int copy_info_to_user(__kernel_fsid_t *fsid, struct fanotify_fh *fh,
-			     const char *name, size_t name_len,
+			     int info_type, const char *name, size_t name_len,
 			     char __user *buf, size_t count)
 {
 	struct fanotify_event_info_fid info = { };
@@ -231,7 +239,7 @@ static int copy_info_to_user(__kernel_fsid_t *fsid, struct fanotify_fh *fh,
 	pr_debug("%s: fh_len=%zu name_len=%zu, info_len=%zu, count=%zu\n",
 		 __func__, fh_len, name_len, info_len, count);
 
-	if (!fh_len || (name && !name_len))
+	if (!fh_len)
 		return 0;
 
 	if (WARN_ON_ONCE(len < sizeof(info) || len > count))
@@ -241,8 +249,21 @@ static int copy_info_to_user(__kernel_fsid_t *fsid, struct fanotify_fh *fh,
 	 * Copy event info fid header followed by variable sized file handle
 	 * and optionally followed by variable sized filename.
 	 */
-	info.hdr.info_type = name_len ? FAN_EVENT_INFO_TYPE_DFID_NAME :
-					FAN_EVENT_INFO_TYPE_FID;
+	switch (info_type) {
+	case FAN_EVENT_INFO_TYPE_FID:
+	case FAN_EVENT_INFO_TYPE_DFID:
+		if (WARN_ON_ONCE(name_len))
+			return -EFAULT;
+		break;
+	case FAN_EVENT_INFO_TYPE_DFID_NAME:
+		if (WARN_ON_ONCE(!name || !name_len))
+			return -EFAULT;
+		break;
+	default:
+		return -EFAULT;
+	}
+
+	info.hdr.info_type = info_type;
 	info.hdr.len = len;
 	info.fsid = *fsid;
 	if (copy_to_user(buf, &info, sizeof(info)))
@@ -305,13 +326,16 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
 {
 	struct fanotify_event_metadata metadata;
 	struct path *path = fanotify_event_path(event);
+	struct fanotify_info *info = fanotify_event_info(event);
+	unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
 	struct file *f = NULL;
 	int ret, fd = FAN_NOFD;
+	int info_type = 0;
 
 	pr_debug("%s: group=%p event=%p\n", __func__, group, event);
 
 	metadata.event_len = FAN_EVENT_METADATA_LEN +
-					fanotify_event_info_len(event);
+				fanotify_event_info_len(fid_mode, event);
 	metadata.metadata_len = FAN_EVENT_METADATA_LEN;
 	metadata.vers = FANOTIFY_METADATA_VERSION;
 	metadata.reserved = 0;
@@ -346,13 +370,13 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
 		fd_install(fd, f);
 
 	/* Event info records order is: dir fid + name, child fid */
-	if (fanotify_event_name_len(event)) {
-		struct fanotify_name_event *fne = FANOTIFY_NE(event);
-
+	if (fanotify_event_dir_fh_len(event)) {
+		info_type = info->name_len ? FAN_EVENT_INFO_TYPE_DFID_NAME :
+					     FAN_EVENT_INFO_TYPE_DFID;
 		ret = copy_info_to_user(fanotify_event_fsid(event),
-					fanotify_event_dir_fh(event),
-					fne->name, fne->name_len,
-					buf, count);
+					fanotify_info_dir_fh(info),
+					info_type, fanotify_info_name(info),
+					info->name_len, buf, count);
 		if (ret < 0)
 			return ret;
 
@@ -361,9 +385,46 @@ static ssize_t copy_event_to_user(struct fsnotify_group *group,
 	}
 
 	if (fanotify_event_object_fh_len(event)) {
+		const char *dot = NULL;
+		int dot_len = 0;
+
+		if (fid_mode == FAN_REPORT_FID || info_type) {
+			/*
+			 * With only group flag FAN_REPORT_FID only type FID is
+			 * reported. Second info record type is always FID.
+			 */
+			info_type = FAN_EVENT_INFO_TYPE_FID;
+		} else if ((fid_mode & FAN_REPORT_NAME) &&
+			   (event->mask & FAN_ONDIR)) {
+			/*
+			 * With group flag FAN_REPORT_NAME, if name was not
+			 * recorded in an event on a directory, report the
+			 * name "." with info type DFID_NAME.
+			 */
+			info_type = FAN_EVENT_INFO_TYPE_DFID_NAME;
+			dot = ".";
+			dot_len = 1;
+		} else if ((event->mask & ALL_FSNOTIFY_DIRENT_EVENTS) ||
+			   (event->mask & FAN_ONDIR)) {
+			/*
+			 * With group flag FAN_REPORT_DIR_FID, a single info
+			 * record has type DFID for directory entry modification
+			 * event and for event on a directory.
+			 */
+			info_type = FAN_EVENT_INFO_TYPE_DFID;
+		} else {
+			/*
+			 * With group flags FAN_REPORT_DIR_FID|FAN_REPORT_FID,
+			 * a single info record has type FID for event on a
+			 * non-directory, when there is no directory to report.
+			 * For example, on FAN_DELETE_SELF event.
+			 */
+			info_type = FAN_EVENT_INFO_TYPE_FID;
+		}
+
 		ret = copy_info_to_user(fanotify_event_fsid(event),
 					fanotify_event_object_fh(event),
-					NULL, 0, buf, count);
+					info_type, dot, dot_len, buf, count);
 		if (ret < 0)
 			return ret;
 
@@ -412,6 +473,11 @@ static ssize_t fanotify_read(struct file *file, char __user *buf,
 
 	add_wait_queue(&group->notification_waitq, &wait);
 	while (1) {
+		/*
+		 * User can supply arbitrarily large buffer. Avoid softlockups
+		 * in case there are lots of available events.
+		 */
+		cond_resched();
 		event = get_one_event(group, count);
 		if (IS_ERR(event)) {
 			ret = PTR_ERR(event);
@@ -651,12 +717,13 @@ static int fanotify_find_path(int dfd, const char __user *filename,
 }
 
 static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
-					    __u32 mask,
-					    unsigned int flags,
-					    int *destroy)
+					    __u32 mask, unsigned int flags,
+					    __u32 umask, int *destroy)
 {
 	__u32 oldmask = 0;
 
+	/* umask bits cannot be removed by user */
+	mask &= ~umask;
 	spin_lock(&fsn_mark->lock);
 	if (!(flags & FAN_MARK_IGNORED_MASK)) {
 		oldmask = fsn_mark->mask;
@@ -664,7 +731,13 @@ static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
 	} else {
 		fsn_mark->ignored_mask &= ~mask;
 	}
-	*destroy = !(fsn_mark->mask | fsn_mark->ignored_mask);
+	/*
+	 * We need to keep the mark around even if remaining mask cannot
+	 * result in any events (e.g. mask == FAN_ONDIR) to support incremenal
+	 * changes to the mask.
+	 * Destroy mark when only umask bits remain.
+	 */
+	*destroy = !((fsn_mark->mask | fsn_mark->ignored_mask) & ~umask);
 	spin_unlock(&fsn_mark->lock);
 
 	return mask & oldmask;
@@ -672,7 +745,7 @@ static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
 
 static int fanotify_remove_mark(struct fsnotify_group *group,
 				fsnotify_connp_t *connp, __u32 mask,
-				unsigned int flags)
+				unsigned int flags, __u32 umask)
 {
 	struct fsnotify_mark *fsn_mark = NULL;
 	__u32 removed;
@@ -686,7 +759,7 @@ static int fanotify_remove_mark(struct fsnotify_group *group,
 	}
 
 	removed = fanotify_mark_remove_from_mask(fsn_mark, mask, flags,
-						 &destroy_mark);
+						 umask, &destroy_mark);
 	if (removed & fsnotify_conn_mask(fsn_mark->connector))
 		fsnotify_recalc_mask(fsn_mark->connector);
 	if (destroy_mark)
@@ -702,25 +775,26 @@ static int fanotify_remove_mark(struct fsnotify_group *group,
 
 static int fanotify_remove_vfsmount_mark(struct fsnotify_group *group,
 					 struct vfsmount *mnt, __u32 mask,
-					 unsigned int flags)
+					 unsigned int flags, __u32 umask)
 {
 	return fanotify_remove_mark(group, &real_mount(mnt)->mnt_fsnotify_marks,
-				    mask, flags);
+				    mask, flags, umask);
 }
 
 static int fanotify_remove_sb_mark(struct fsnotify_group *group,
-				      struct super_block *sb, __u32 mask,
-				      unsigned int flags)
+				   struct super_block *sb, __u32 mask,
+				   unsigned int flags, __u32 umask)
 {
-	return fanotify_remove_mark(group, &sb->s_fsnotify_marks, mask, flags);
+	return fanotify_remove_mark(group, &sb->s_fsnotify_marks, mask,
+				    flags, umask);
 }
 
 static int fanotify_remove_inode_mark(struct fsnotify_group *group,
 				      struct inode *inode, __u32 mask,
-				      unsigned int flags)
+				      unsigned int flags, __u32 umask)
 {
 	return fanotify_remove_mark(group, &inode->i_fsnotify_marks, mask,
-				    flags);
+				    flags, umask);
 }
 
 static __u32 fanotify_mark_add_to_mask(struct fsnotify_mark *fsn_mark,
@@ -831,13 +905,28 @@ static int fanotify_add_inode_mark(struct fsnotify_group *group,
 				 FSNOTIFY_OBJ_TYPE_INODE, mask, flags, fsid);
 }
 
+static struct fsnotify_event *fanotify_alloc_overflow_event(void)
+{
+	struct fanotify_event *oevent;
+
+	oevent = kmalloc(sizeof(*oevent), GFP_KERNEL_ACCOUNT);
+	if (!oevent)
+		return NULL;
+
+	fanotify_init_event(oevent, 0, FS_Q_OVERFLOW);
+	oevent->type = FANOTIFY_EVENT_TYPE_OVERFLOW;
+
+	return &oevent->fse;
+}
+
 /* fanotify syscalls */
 SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 {
 	struct fsnotify_group *group;
 	int f_flags, fd;
 	struct user_struct *user;
-	struct fanotify_event *oevent;
+	unsigned int fid_mode = flags & FANOTIFY_FID_BITS;
+	unsigned int class = flags & FANOTIFY_CLASS_BITS;
 
 	pr_debug("%s: flags=%x event_f_flags=%x\n",
 		 __func__, flags, event_f_flags);
@@ -864,8 +953,14 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 		return -EINVAL;
 	}
 
-	if ((flags & FAN_REPORT_FID) &&
-	    (flags & FANOTIFY_CLASS_BITS) != FAN_CLASS_NOTIF)
+	if (fid_mode && class != FAN_CLASS_NOTIF)
+		return -EINVAL;
+
+	/*
+	 * Child name is reported with parent fid so requires dir fid.
+	 * We can report both child fid and dir fid with or without name.
+	 */
+	if ((fid_mode & FAN_REPORT_NAME) && !(fid_mode & FAN_REPORT_DIR_FID))
 		return -EINVAL;
 
 	user = get_current_user();
@@ -892,20 +987,18 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
 	atomic_inc(&user->fanotify_listeners);
 	group->memcg = get_mem_cgroup_from_mm(current->mm);
 
-	oevent = fanotify_alloc_event(group, NULL, FS_Q_OVERFLOW, NULL,
-				      FSNOTIFY_EVENT_NONE, NULL, NULL);
-	if (unlikely(!oevent)) {
+	group->overflow_event = fanotify_alloc_overflow_event();
+	if (unlikely(!group->overflow_event)) {
 		fd = -ENOMEM;
 		goto out_destroy_group;
 	}
-	group->overflow_event = &oevent->fse;
 
 	if (force_o_largefile())
 		event_f_flags |= O_LARGEFILE;
 	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 & FANOTIFY_CLASS_BITS) {
+	switch (class) {
 	case FAN_CLASS_NOTIF:
 		group->priority = FS_PRIO_0;
 		break;
@@ -1024,7 +1117,9 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	__kernel_fsid_t __fsid, *fsid = NULL;
 	u32 valid_mask = FANOTIFY_EVENTS | FANOTIFY_EVENT_FLAGS;
 	unsigned int mark_type = flags & FANOTIFY_MARK_TYPE_BITS;
-	unsigned int obj_type;
+	bool ignored = flags & FAN_MARK_IGNORED_MASK;
+	unsigned int obj_type, fid_mode;
+	u32 umask = 0;
 	int ret;
 
 	pr_debug("%s: fanotify_fd=%d flags=%x dfd=%d pathname=%p mask=%llx\n",
@@ -1071,6 +1166,10 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	if (mask & ~valid_mask)
 		return -EINVAL;
 
+	/* Event flags (ONDIR, ON_CHILD) are meaningless in ignored mask */
+	if (ignored)
+		mask &= ~FANOTIFY_EVENT_FLAGS;
+
 	f = fdget(fanotify_fd);
 	if (unlikely(!f.file))
 		return -EBADF;
@@ -1097,9 +1196,9 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	 * inode events are not supported on a mount mark, because they do not
 	 * carry enough information (i.e. path) to be filtered by mount point.
 	 */
+	fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
 	if (mask & FANOTIFY_INODE_EVENTS &&
-	    (!FAN_GROUP_FLAG(group, FAN_REPORT_FID) ||
-	     mark_type == FAN_MARK_MOUNT))
+	    (!fid_mode || mark_type == FAN_MARK_MOUNT))
 		goto fput_and_out;
 
 	if (flags & FAN_MARK_FLUSH) {
@@ -1124,7 +1223,7 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 			goto path_put_and_out;
 	}
 
-	if (FAN_GROUP_FLAG(group, FAN_REPORT_FID)) {
+	if (fid_mode) {
 		ret = fanotify_test_fid(&path, &__fsid);
 		if (ret)
 			goto path_put_and_out;
@@ -1138,6 +1237,19 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	else
 		mnt = path.mnt;
 
+	/* Mask out FAN_EVENT_ON_CHILD flag for sb/mount/non-dir marks */
+	if (mnt || !S_ISDIR(inode->i_mode)) {
+		mask &= ~FAN_EVENT_ON_CHILD;
+		umask = FAN_EVENT_ON_CHILD;
+		/*
+		 * If group needs to report parent fid, register for getting
+		 * events with parent/name info for non-directory.
+		 */
+		if ((fid_mode & FAN_REPORT_DIR_FID) &&
+		    (flags & FAN_MARK_ADD) && !ignored)
+			mask |= FAN_EVENT_ON_CHILD;
+	}
+
 	/* create/update an inode mark */
 	switch (flags & (FAN_MARK_ADD | FAN_MARK_REMOVE)) {
 	case FAN_MARK_ADD:
@@ -1154,13 +1266,13 @@ static int do_fanotify_mark(int fanotify_fd, unsigned int flags, __u64 mask,
 	case FAN_MARK_REMOVE:
 		if (mark_type == FAN_MARK_MOUNT)
 			ret = fanotify_remove_vfsmount_mark(group, mnt, mask,
-							    flags);
+							    flags, umask);
 		else if (mark_type == FAN_MARK_FILESYSTEM)
 			ret = fanotify_remove_sb_mark(group, mnt->mnt_sb, mask,
-						      flags);
+						      flags, umask);
 		else
 			ret = fanotify_remove_inode_mark(group, inode, mask,
-							 flags);
+							 flags, umask);
 		break;
 	default:
 		ret = -EINVAL;
@@ -1203,7 +1315,7 @@ COMPAT_SYSCALL_DEFINE6(fanotify_mark,
  */
 static int __init fanotify_user_setup(void)
 {
-	BUILD_BUG_ON(HWEIGHT32(FANOTIFY_INIT_FLAGS) != 8);
+	BUILD_BUG_ON(HWEIGHT32(FANOTIFY_INIT_FLAGS) != 10);
 	BUILD_BUG_ON(HWEIGHT32(FANOTIFY_MARK_FLAGS) != 9);
 
 	fanotify_mark_cache = KMEM_CACHE(fsnotify_mark,

fs/notify/fsnotify.c

@@ -74,7 +74,7 @@ static void fsnotify_unmount_inodes(struct super_block *sb)
 			iput(iput_inode);
 
 		/* for each watch, send FS_UNMOUNT and then remove it */
-		fsnotify(inode, FS_UNMOUNT, inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+		fsnotify_inode(inode, FS_UNMOUNT);
 
 		fsnotify_inode_delete(inode);
 
@@ -142,45 +142,140 @@ void __fsnotify_update_child_dentry_flags(struct inode *inode)
 	spin_unlock(&inode->i_lock);
 }
 
-/* Notify this dentry's parent about a child's events. */
-int fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
-		    int data_type)
+/* Are inode/sb/mount interested in parent and name info with this event? */
+static bool fsnotify_event_needs_parent(struct inode *inode, struct mount *mnt,
+					__u32 mask)
 {
+	__u32 marks_mask = 0;
+
+	/* We only send parent/name to inode/sb/mount for events on non-dir */
+	if (mask & FS_ISDIR)
+		return false;
+
+	/* Did either inode/sb/mount subscribe for events with parent/name? */
+	marks_mask |= fsnotify_parent_needed_mask(inode->i_fsnotify_mask);
+	marks_mask |= fsnotify_parent_needed_mask(inode->i_sb->s_fsnotify_mask);
+	if (mnt)
+		marks_mask |= fsnotify_parent_needed_mask(mnt->mnt_fsnotify_mask);
+
+	/* Did they subscribe for this event with parent/name info? */
+	return mask & marks_mask;
+}
+
+/*
+ * Notify this dentry's parent about a child's events with child name info
+ * if parent is watching or if inode/sb/mount are interested in events with
+ * parent and name info.
+ *
+ * Notify only the child without name info if parent is not watching and
+ * inode/sb/mount are not interested in events with parent and name info.
+ */
+int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data,
+		      int data_type)
+{
+	const struct path *path = fsnotify_data_path(data, data_type);
+	struct mount *mnt = path ? real_mount(path->mnt) : NULL;
+	struct inode *inode = d_inode(dentry);
 	struct dentry *parent;
-	struct inode *p_inode;
+	bool parent_watched = dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED;
+	__u32 p_mask;
+	struct inode *p_inode = NULL;
+	struct name_snapshot name;
+	struct qstr *file_name = NULL;
 	int ret = 0;
 
-	if (!(dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED))
+	/*
+	 * Do inode/sb/mount care about parent and name info on non-dir?
+	 * Do they care about any event at all?
+	 */
+	if (!inode->i_fsnotify_marks && !inode->i_sb->s_fsnotify_marks &&
+	    (!mnt || !mnt->mnt_fsnotify_marks) && !parent_watched)
 		return 0;
 
+	parent = NULL;
+	if (!parent_watched && !fsnotify_event_needs_parent(inode, mnt, mask))
+		goto notify;
+
+	/* Does parent inode care about events on children? */
 	parent = dget_parent(dentry);
 	p_inode = parent->d_inode;
-
-	if (unlikely(!fsnotify_inode_watches_children(p_inode))) {
+	p_mask = fsnotify_inode_watches_children(p_inode);
+	if (unlikely(parent_watched && !p_mask))
 		__fsnotify_update_child_dentry_flags(p_inode);
-	} else if (p_inode->i_fsnotify_mask & mask & ALL_FSNOTIFY_EVENTS) {
-		struct name_snapshot name;
 
-		/* 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;
+	/*
+	 * Include parent/name in notification either if some notification
+	 * groups require parent info (!parent_watched case) or the parent is
+	 * interested in this event.
+	 */
+	if (!parent_watched || (mask & p_mask & ALL_FSNOTIFY_EVENTS)) {
+		/* When notifying parent, child should be passed as data */
+		WARN_ON_ONCE(inode != fsnotify_data_inode(data, data_type));
 
+		/* Notify both parent and child with child name info */
 		take_dentry_name_snapshot(&name, dentry);
-		ret = fsnotify(p_inode, mask, data, data_type, &name.name, 0);
-		release_dentry_name_snapshot(&name);
+		file_name = &name.name;
+		if (parent_watched)
+			mask |= FS_EVENT_ON_CHILD;
 	}
 
+notify:
+	ret = fsnotify(mask, data, data_type, p_inode, file_name, inode, 0);
+
+	if (file_name)
+		release_dentry_name_snapshot(&name);
 	dput(parent);
 
 	return ret;
 }
-EXPORT_SYMBOL_GPL(fsnotify_parent);
+EXPORT_SYMBOL_GPL(__fsnotify_parent);
+
+static int fsnotify_handle_event(struct fsnotify_group *group, __u32 mask,
+				 const void *data, int data_type,
+				 struct inode *dir, const struct qstr *name,
+				 u32 cookie, struct fsnotify_iter_info *iter_info)
+{
+	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
+	struct fsnotify_mark *child_mark = fsnotify_iter_child_mark(iter_info);
+	struct inode *inode = fsnotify_data_inode(data, data_type);
+	const struct fsnotify_ops *ops = group->ops;
+	int ret;
+
+	if (WARN_ON_ONCE(!ops->handle_inode_event))
+		return 0;
+
+	if (WARN_ON_ONCE(fsnotify_iter_sb_mark(iter_info)) ||
+	    WARN_ON_ONCE(fsnotify_iter_vfsmount_mark(iter_info)))
+		return 0;
+
+	/*
+	 * An event can be sent on child mark iterator instead of inode mark
+	 * iterator because of other groups that have interest of this inode
+	 * and have marks on both parent and child.  We can simplify this case.
+	 */
+	if (!inode_mark) {
+		inode_mark = child_mark;
+		child_mark = NULL;
+		dir = NULL;
+		name = NULL;
+	}
+
+	ret = ops->handle_inode_event(inode_mark, mask, inode, dir, name);
+	if (ret || !child_mark)
+		return ret;
+
+	/*
+	 * Some events can be sent on both parent dir and child marks
+	 * (e.g. FS_ATTRIB).  If both parent dir and child are watching,
+	 * report the event once to parent dir with name and once to child
+	 * without name.
+	 */
+	return ops->handle_inode_event(child_mark, mask, inode, NULL, NULL);
+}
 
-static int send_to_group(struct inode *to_tell,
-			 __u32 mask, const void *data,
-			 int data_is, u32 cookie,
-			 const struct qstr *file_name,
-			 struct fsnotify_iter_info *iter_info)
+static int send_to_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)
 {
 	struct fsnotify_group *group = NULL;
 	__u32 test_mask = (mask & ALL_FSNOTIFY_EVENTS);
@@ -216,16 +311,20 @@ static int send_to_group(struct inode *to_tell,
 		}
 	}
 
-	pr_debug("%s: group=%p to_tell=%p mask=%x marks_mask=%x marks_ignored_mask=%x"
-		 " data=%p data_is=%d cookie=%d\n",
-		 __func__, group, to_tell, mask, marks_mask, marks_ignored_mask,
-		 data, data_is, cookie);
+	pr_debug("%s: group=%p mask=%x marks_mask=%x marks_ignored_mask=%x data=%p data_type=%d dir=%p cookie=%d\n",
+		 __func__, group, mask, marks_mask, marks_ignored_mask,
+		 data, data_type, dir, cookie);
 
 	if (!(test_mask & marks_mask & ~marks_ignored_mask))
 		return 0;
 
-	return group->ops->handle_event(group, to_tell, mask, data, data_is,
-					file_name, cookie, iter_info);
+	if (group->ops->handle_event) {
+		return group->ops->handle_event(group, mask, data, data_type, dir,
+						file_name, cookie, iter_info);
+	}
+
+	return fsnotify_handle_event(group, mask, data, data_type, dir,
+				     file_name, cookie, iter_info);
 }
 
 static struct fsnotify_mark *fsnotify_first_mark(struct fsnotify_mark_connector **connp)
@@ -303,29 +402,51 @@ static void fsnotify_iter_next(struct fsnotify_iter_info *iter_info)
 }
 
 /*
- * 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
- * notification event in whatever means they feel necessary.
+ * fsnotify - 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 notification event
+ * in whatever means they feel necessary.
+ *
+ * @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
+ * @inode:	optional inode associated with event -
+ *		either @dir or @inode must be non-NULL.
+ *		if both are non-NULL event may be reported to both.
+ * @cookie:	inotify rename cookie
  */
-int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
-	     const struct qstr *file_name, u32 cookie)
+int fsnotify(__u32 mask, const void *data, int data_type, struct inode *dir,
+	     const struct qstr *file_name, struct inode *inode, u32 cookie)
 {
-	const struct path *path = fsnotify_data_path(data, data_is);
+	const struct path *path = fsnotify_data_path(data, data_type);
 	struct fsnotify_iter_info iter_info = {};
-	struct super_block *sb = to_tell->i_sb;
+	struct super_block *sb;
 	struct mount *mnt = NULL;
-	__u32 mnt_or_sb_mask = sb->s_fsnotify_mask;
+	struct inode *child = NULL;
 	int ret = 0;
-	__u32 test_mask = (mask & ALL_FSNOTIFY_EVENTS);
+	__u32 test_mask, marks_mask;
 
-	if (path) {
+	if (path)
 		mnt = real_mount(path->mnt);
-		mnt_or_sb_mask |= mnt->mnt_fsnotify_mask;
+
+	if (!inode) {
+		/* Dirent event - report on TYPE_INODE to dir */
+		inode = dir;
+	} else if (mask & FS_EVENT_ON_CHILD) {
+		/*
+		 * Event on child - report on TYPE_INODE to dir if it is
+		 * watching children and on TYPE_CHILD to child.
+		 */
+		child = inode;
+		inode = dir;
 	}
-	/* An event "on child" is not intended for a mount/sb mark */
-	if (mask & FS_EVENT_ON_CHILD)
-		mnt_or_sb_mask = 0;
+	sb = inode->i_sb;
 
 	/*
 	 * Optimization: srcu_read_lock() has a memory barrier which can
@@ -334,28 +455,45 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
 	 * SRCU because we have no references to any objects and do not
 	 * need SRCU to keep them "alive".
 	 */
-	if (!to_tell->i_fsnotify_marks && !sb->s_fsnotify_marks &&
-	    (!mnt || !mnt->mnt_fsnotify_marks))
+	if (!sb->s_fsnotify_marks &&
+	    (!mnt || !mnt->mnt_fsnotify_marks) &&
+	    (!inode || !inode->i_fsnotify_marks) &&
+	    (!child || !child->i_fsnotify_marks))
 		return 0;
+
+	marks_mask = sb->s_fsnotify_mask;
+	if (mnt)
+		marks_mask |= mnt->mnt_fsnotify_mask;
+	if (inode)
+		marks_mask |= inode->i_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.
 	 */
-	if (!(mask & FS_MODIFY) &&
-	    !(test_mask & (to_tell->i_fsnotify_mask | mnt_or_sb_mask)))
+	test_mask = (mask & ALL_FSNOTIFY_EVENTS);
+	if (!(mask & FS_MODIFY) && !(test_mask & marks_mask))
 		return 0;
 
 	iter_info.srcu_idx = srcu_read_lock(&fsnotify_mark_srcu);
 
-	iter_info.marks[FSNOTIFY_OBJ_TYPE_INODE] =
-		fsnotify_first_mark(&to_tell->i_fsnotify_marks);
 	iter_info.marks[FSNOTIFY_OBJ_TYPE_SB] =
 		fsnotify_first_mark(&sb->s_fsnotify_marks);
 	if (mnt) {
 		iter_info.marks[FSNOTIFY_OBJ_TYPE_VFSMOUNT] =
 			fsnotify_first_mark(&mnt->mnt_fsnotify_marks);
 	}
+	if (inode) {
+		iter_info.marks[FSNOTIFY_OBJ_TYPE_INODE] =
+			fsnotify_first_mark(&inode->i_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
@@ -363,8 +501,8 @@ int fsnotify(struct inode *to_tell, __u32 mask, const void *data, int data_is,
 	 * That's why this traversal is so complicated...
 	 */
 	while (fsnotify_iter_select_report_types(&iter_info)) {
-		ret = send_to_group(to_tell, mask, data, data_is, cookie,
-				    file_name, &iter_info);
+		ret = send_to_group(mask, data, data_type, dir, file_name,
+				    cookie, &iter_info);
 
 		if (ret && (mask & ALL_FSNOTIFY_PERM_EVENTS))
 			goto out;
@@ -383,7 +521,7 @@ static __init int fsnotify_init(void)
 {
 	int ret;
 
-	BUILD_BUG_ON(HWEIGHT32(ALL_FSNOTIFY_BITS) != 26);
+	BUILD_BUG_ON(HWEIGHT32(ALL_FSNOTIFY_BITS) != 25);
 
 	ret = init_srcu_struct(&fsnotify_mark_srcu);
 	if (ret)

fs/notify/inotify/inotify.h

@@ -24,9 +24,9 @@ static inline struct inotify_event_info *INOTIFY_E(struct fsnotify_event *fse)
 
 extern void inotify_ignored_and_remove_idr(struct fsnotify_mark *fsn_mark,
 					   struct fsnotify_group *group);
-extern int inotify_handle_event(struct fsnotify_group *group,
-				struct inode *inode,
-				u32 mask, const void *data, int data_type,
+extern int inotify_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);
 

fs/notify/inotify/inotify_fsnotify.c

@@ -39,7 +39,7 @@ static bool event_compare(struct fsnotify_event *old_fsn,
 	if (old->mask & FS_IN_IGNORED)
 		return false;
 	if ((old->mask == new->mask) &&
-	    (old_fsn->objectid == new_fsn->objectid) &&
+	    (old->wd == new->wd) &&
 	    (old->name_len == new->name_len) &&
 	    (!old->name_len || !strcmp(old->name, new->name)))
 		return true;
@@ -55,14 +55,11 @@ static int inotify_merge(struct list_head *list,
 	return event_compare(last_event, event);
 }
 
-int inotify_handle_event(struct fsnotify_group *group,
-			 struct inode *inode,
-			 u32 mask, const void *data, int data_type,
-			 const struct qstr *file_name, u32 cookie,
-			 struct fsnotify_iter_info *iter_info)
+static int inotify_one_event(struct fsnotify_group *group, u32 mask,
+			     struct fsnotify_mark *inode_mark,
+			     const struct path *path,
+			     const struct qstr *file_name, u32 cookie)
 {
-	const struct path *path = fsnotify_data_path(data, data_type);
-	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
 	struct inotify_inode_mark *i_mark;
 	struct inotify_event_info *event;
 	struct fsnotify_event *fsn_event;
@@ -70,9 +67,6 @@ int inotify_handle_event(struct fsnotify_group *group,
 	int len = 0;
 	int alloc_len = sizeof(struct inotify_event_info);
 
-	if (WARN_ON(fsnotify_iter_vfsmount_mark(iter_info)))
-		return 0;
-
 	if ((inode_mark->mask & FS_EXCL_UNLINK) &&
 	    path && d_unlinked(path->dentry))
 		return 0;
@@ -82,7 +76,7 @@ int inotify_handle_event(struct fsnotify_group *group,
 		alloc_len += len + 1;
 	}
 
-	pr_debug("%s: group=%p inode=%p mask=%x\n", __func__, group, inode,
+	pr_debug("%s: group=%p mark=%p mask=%x\n", __func__, group, inode_mark,
 		 mask);
 
 	i_mark = container_of(inode_mark, struct inotify_inode_mark,
@@ -116,7 +110,7 @@ int inotify_handle_event(struct fsnotify_group *group,
 		mask &= ~IN_ISDIR;
 
 	fsn_event = &event->fse;
-	fsnotify_init_event(fsn_event, (unsigned long)inode);
+	fsnotify_init_event(fsn_event, 0);
 	event->mask = mask;
 	event->wd = i_mark->wd;
 	event->sync_cookie = cookie;
@@ -136,6 +130,37 @@ int inotify_handle_event(struct fsnotify_group *group,
 	return 0;
 }
 
+int inotify_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)
+{
+	const struct path *path = fsnotify_data_path(data, data_type);
+	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
+	struct fsnotify_mark *child_mark = fsnotify_iter_child_mark(iter_info);
+	int ret = 0;
+
+	if (WARN_ON(fsnotify_iter_vfsmount_mark(iter_info)))
+		return 0;
+
+	/*
+	 * Some events cannot be sent on both parent and child marks
+	 * (e.g. IN_CREATE).  Those events are always sent on inode_mark.
+	 * For events that are possible on both parent and child (e.g. IN_OPEN),
+	 * event is sent on inode_mark with name if the parent is watching and
+	 * is sent on child_mark without name if child is watching.
+	 * If both parent and child are watching, report the event with child's
+	 * name here and report another event without child's name below.
+	 */
+	if (inode_mark)
+		ret = inotify_one_event(group, mask, inode_mark, path,
+					file_name, cookie);
+	if (ret || !child_mark)
+		return ret;
+
+	return inotify_one_event(group, mask, child_mark, path, NULL, 0);
+}
+
 static void inotify_freeing_mark(struct fsnotify_mark *fsn_mark, struct fsnotify_group *group)
 {
 	inotify_ignored_and_remove_idr(fsn_mark, group);

fs/notify/inotify/inotify_user.c

@@ -75,15 +75,17 @@ struct ctl_table inotify_table[] = {
 };
 #endif /* CONFIG_SYSCTL */
 
-static inline __u32 inotify_arg_to_mask(u32 arg)
+static inline __u32 inotify_arg_to_mask(struct inode *inode, u32 arg)
 {
 	__u32 mask;
 
 	/*
-	 * everything should accept their own ignored, cares about children,
-	 * and should receive events when the inode is unmounted
+	 * Everything should accept their own ignored and should receive events
+	 * when the inode is unmounted.  All directories care about children.
 	 */
-	mask = (FS_IN_IGNORED | FS_EVENT_ON_CHILD | FS_UNMOUNT);
+	mask = (FS_IN_IGNORED | FS_UNMOUNT);
+	if (S_ISDIR(inode->i_mode))
+		mask |= FS_EVENT_ON_CHILD;
 
 	/* mask off the flags used to open the fd */
 	mask |= (arg & (IN_ALL_EVENTS | IN_ONESHOT | IN_EXCL_UNLINK));
@@ -490,8 +492,8 @@ void inotify_ignored_and_remove_idr(struct fsnotify_mark *fsn_mark,
 					   fsn_mark);
 
 	/* Queue ignore event for the watch */
-	inotify_handle_event(group, NULL, FS_IN_IGNORED, NULL,
-			     FSNOTIFY_EVENT_NONE, NULL, 0, &iter_info);
+	inotify_handle_event(group, FS_IN_IGNORED, NULL, FSNOTIFY_EVENT_NONE,
+			     NULL, NULL, 0, &iter_info);
 
 	i_mark = container_of(fsn_mark, struct inotify_inode_mark, fsn_mark);
 	/* remove this mark from the idr */
@@ -512,7 +514,7 @@ static int inotify_update_existing_watch(struct fsnotify_group *group,
 	int create = (arg & IN_MASK_CREATE);
 	int ret;
 
-	mask = inotify_arg_to_mask(arg);
+	mask = inotify_arg_to_mask(inode, arg);
 
 	fsn_mark = fsnotify_find_mark(&inode->i_fsnotify_marks, group);
 	if (!fsn_mark)
@@ -565,7 +567,7 @@ static int inotify_new_watch(struct fsnotify_group *group,
 	struct idr *idr = &group->inotify_data.idr;
 	spinlock_t *idr_lock = &group->inotify_data.idr_lock;
 
-	mask = inotify_arg_to_mask(arg);
+	mask = inotify_arg_to_mask(inode, arg);
 
 	tmp_i_mark = kmem_cache_alloc(inotify_inode_mark_cachep, GFP_KERNEL);
 	if (unlikely(!tmp_i_mark))

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)
 

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);
 }
 

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;

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;

kernel/audit_fsnotify.c

@@ -36,7 +36,7 @@ static struct fsnotify_group *audit_fsnotify_group;
 
 /* fsnotify events we care about. */
 #define AUDIT_FS_EVENTS (FS_MOVE | FS_CREATE | FS_DELETE | FS_DELETE_SELF |\
-			 FS_MOVE_SELF | FS_EVENT_ON_CHILD)
+			 FS_MOVE_SELF)
 
 static void audit_fsnotify_mark_free(struct audit_fsnotify_mark *audit_mark)
 {
@@ -152,35 +152,31 @@ static void audit_autoremove_mark_rule(struct audit_fsnotify_mark *audit_mark)
 }
 
 /* Update mark data in audit rules based on fsnotify events. */
-static int audit_mark_handle_event(struct fsnotify_group *group,
-				    struct inode *to_tell,
-				    u32 mask, const void *data, int data_type,
-				    const struct qstr *dname, u32 cookie,
-				    struct fsnotify_iter_info *iter_info)
+static int audit_mark_handle_event(struct fsnotify_mark *inode_mark, u32 mask,
+				   struct inode *inode, struct inode *dir,
+				   const struct qstr *dname)
 {
-	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
 	struct audit_fsnotify_mark *audit_mark;
-	const struct inode *inode = fsnotify_data_inode(data, data_type);
 
 	audit_mark = container_of(inode_mark, struct audit_fsnotify_mark, mark);
 
-	BUG_ON(group != audit_fsnotify_group);
-
-	if (WARN_ON(!inode))
+	if (WARN_ON_ONCE(inode_mark->group != audit_fsnotify_group) ||
+	    WARN_ON_ONCE(!inode))
 		return 0;
 
 	if (mask & (FS_CREATE|FS_MOVED_TO|FS_DELETE|FS_MOVED_FROM)) {
 		if (audit_compare_dname_path(dname, audit_mark->path, AUDIT_NAME_FULL))
 			return 0;
 		audit_update_mark(audit_mark, inode);
-	} else if (mask & (FS_DELETE_SELF|FS_UNMOUNT|FS_MOVE_SELF))
+	} else if (mask & (FS_DELETE_SELF|FS_UNMOUNT|FS_MOVE_SELF)) {
 		audit_autoremove_mark_rule(audit_mark);
+	}
 
 	return 0;
 }
 
 static const struct fsnotify_ops audit_mark_fsnotify_ops = {
-	.handle_event =	audit_mark_handle_event,
+	.handle_inode_event = audit_mark_handle_event,
 	.free_mark = audit_fsnotify_free_mark,
 };
 

kernel/audit_tree.c

@@ -1035,11 +1035,9 @@ static void evict_chunk(struct audit_chunk *chunk)
 		audit_schedule_prune();
 }
 
-static int audit_tree_handle_event(struct fsnotify_group *group,
-				   struct inode *to_tell,
-				   u32 mask, const void *data, int data_type,
-				   const struct qstr *file_name, u32 cookie,
-				   struct fsnotify_iter_info *iter_info)
+static int audit_tree_handle_event(struct fsnotify_mark *mark, u32 mask,
+				   struct inode *inode, struct inode *dir,
+				   const struct qstr *file_name)
 {
 	return 0;
 }
@@ -1068,7 +1066,7 @@ static void audit_tree_freeing_mark(struct fsnotify_mark *mark,
 }
 
 static const struct fsnotify_ops audit_tree_ops = {
-	.handle_event = audit_tree_handle_event,
+	.handle_inode_event = audit_tree_handle_event,
 	.freeing_mark = audit_tree_freeing_mark,
 	.free_mark = audit_tree_destroy_watch,
 };

kernel/audit_watch.c

@@ -53,7 +53,7 @@ static struct fsnotify_group *audit_watch_group;
 
 /* fsnotify events we care about. */
 #define AUDIT_FS_WATCH (FS_MOVE | FS_CREATE | FS_DELETE | FS_DELETE_SELF |\
-			FS_MOVE_SELF | FS_EVENT_ON_CHILD | FS_UNMOUNT)
+			FS_MOVE_SELF | FS_UNMOUNT)
 
 static void audit_free_parent(struct audit_parent *parent)
 {
@@ -464,20 +464,17 @@ void audit_remove_watch_rule(struct audit_krule *krule)
 }
 
 /* Update watch data in audit rules based on fsnotify events. */
-static int audit_watch_handle_event(struct fsnotify_group *group,
-				    struct inode *to_tell,
-				    u32 mask, const void *data, int data_type,
-				    const struct qstr *dname, u32 cookie,
-				    struct fsnotify_iter_info *iter_info)
+static int audit_watch_handle_event(struct fsnotify_mark *inode_mark, u32 mask,
+				    struct inode *inode, struct inode *dir,
+				    const struct qstr *dname)
 {
-	struct fsnotify_mark *inode_mark = fsnotify_iter_inode_mark(iter_info);
-	const struct inode *inode = fsnotify_data_inode(data, data_type);
 	struct audit_parent *parent;
 
 	parent = container_of(inode_mark, struct audit_parent, mark);
 
-	BUG_ON(group != audit_watch_group);
-	WARN_ON(!inode);
+	if (WARN_ON_ONCE(inode_mark->group != audit_watch_group) ||
+	    WARN_ON_ONCE(!inode))
+		return 0;
 
 	if (mask & (FS_CREATE|FS_MOVED_TO) && inode)
 		audit_update_watch(parent, dname, inode->i_sb->s_dev, inode->i_ino, 0);
@@ -490,7 +487,7 @@ static int audit_watch_handle_event(struct fsnotify_group *group,
 }
 
 static const struct fsnotify_ops audit_watch_fsnotify_ops = {
-	.handle_event = 	audit_watch_handle_event,
+	.handle_inode_event =	audit_watch_handle_event,
 	.free_mark =		audit_watch_free_mark,
 };
 

kernel/trace/trace.c

@@ -1543,8 +1543,7 @@ static void latency_fsnotify_workfn(struct work_struct *work)
 {
 	struct trace_array *tr = container_of(work, struct trace_array,
 					      fsnotify_work);
-	fsnotify(tr->d_max_latency->d_inode, FS_MODIFY,
-		 tr->d_max_latency->d_inode, FSNOTIFY_EVENT_INODE, NULL, 0);
+	fsnotify_inode(tr->d_max_latency->d_inode, FS_MODIFY);
 }
 
 static void latency_fsnotify_workfn_irq(struct irq_work *iwork)