1 files changed, 172 insertions, 74 deletions

diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c
index b6091775aa6e..a2a15bc4df28 100644
--- a/fs/notify/fanotify/fanotify.c
+++ b/fs/notify/fanotify/fanotify.c
@@ -18,7 +18,7 @@
 
 #include "fanotify.h"
 
-static bool fanotify_path_equal(struct path *p1, struct path *p2)
+static bool fanotify_path_equal(const struct path *p1, const struct path *p2)
 {
 	return p1->mnt == p2->mnt && p1->dentry == p2->dentry;
 }
@@ -76,8 +76,10 @@ static bool fanotify_info_equal(struct fanotify_info *info1,
 				struct fanotify_info *info2)
 {
 	if (info1->dir_fh_totlen != info2->dir_fh_totlen ||
+	    info1->dir2_fh_totlen != info2->dir2_fh_totlen ||
 	    info1->file_fh_totlen != info2->file_fh_totlen ||
-	    info1->name_len != info2->name_len)
+	    info1->name_len != info2->name_len ||
+	    info1->name2_len != info2->name2_len)
 		return false;
 
 	if (info1->dir_fh_totlen &&
@@ -85,14 +87,24 @@ static bool fanotify_info_equal(struct fanotify_info *info1,
 			       fanotify_info_dir_fh(info2)))
 		return false;
 
+	if (info1->dir2_fh_totlen &&
+	    !fanotify_fh_equal(fanotify_info_dir2_fh(info1),
+			       fanotify_info_dir2_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);
+	if (info1->name_len &&
+	    memcmp(fanotify_info_name(info1), fanotify_info_name(info2),
+		   info1->name_len))
+		return false;
+
+	return !info1->name2_len ||
+		!memcmp(fanotify_info_name2(info1), fanotify_info_name2(info2),
+			info1->name2_len);
 }
 
 static bool fanotify_name_event_equal(struct fanotify_name_event *fne1,
@@ -141,6 +153,13 @@ static bool fanotify_should_merge(struct fanotify_event *old,
 	if ((old->mask & FS_ISDIR) != (new->mask & FS_ISDIR))
 		return false;
 
+	/*
+	 * FAN_RENAME event is reported with special info record types,
+	 * so we cannot merge it with other events.
+	 */
+	if ((old->mask & FAN_RENAME) != (new->mask & FAN_RENAME))
+		return false;
+
 	switch (old->type) {
 	case FANOTIFY_EVENT_TYPE_PATH:
 		return fanotify_path_equal(fanotify_event_path(old),
@@ -272,15 +291,17 @@ out:
  */
 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, struct inode *dir)
+				     u32 *match_mask, u32 event_mask,
+				     const void *data, int data_type,
+				     struct inode *dir)
 {
-	__u32 marks_mask = 0, marks_ignored_mask = 0;
+	__u32 marks_mask = 0, marks_ignore_mask = 0;
 	__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;
+	bool ondir = event_mask & FAN_ONDIR;
 	int type;
 
 	pr_debug("%s: report_mask=%x mask=%x data=%p data_type=%d\n",
@@ -295,37 +316,30 @@ static u32 fanotify_group_event_mask(struct fsnotify_group *group,
 			return 0;
 	} else if (!(fid_mode & FAN_REPORT_FID)) {
 		/* Do we have a directory inode to report? */
-		if (!dir && !(event_mask & FS_ISDIR))
+		if (!dir && !ondir)
 			return 0;
 	}
 
-	fsnotify_foreach_obj_type(type) {
-		if (!fsnotify_iter_should_report_type(iter_info, type))
-			continue;
-		mark = iter_info->marks[type];
-
-		/* Apply ignore mask regardless of ISDIR and ON_CHILD flags */
-		marks_ignored_mask |= mark->ignored_mask;
-
+	fsnotify_foreach_iter_mark_type(iter_info, mark, type) {
 		/*
-		 * If the event is on dir and this mark doesn't care about
-		 * events on dir, don't send it!
+		 * Apply ignore mask depending on event flags in ignore mask.
 		 */
-		if (event_mask & FS_ISDIR && !(mark->mask & FS_ISDIR))
-			continue;
+		marks_ignore_mask |=
+			fsnotify_effective_ignore_mask(mark, ondir, type);
 
 		/*
-		 * If the event is on a child and this mark is on a parent not
-		 * watching children, don't send it!
+		 * Send the event depending on event flags in mark mask.
 		 */
-		if (type == FSNOTIFY_OBJ_TYPE_PARENT &&
-		    !(mark->mask & FS_EVENT_ON_CHILD))
+		if (!fsnotify_mask_applicable(mark->mask, ondir, type))
 			continue;
 
 		marks_mask |= mark->mask;
+
+		/* Record the mark types of this group that matched the event */
+		*match_mask |= 1U << type;
 	}
 
-	test_mask = event_mask & marks_mask & ~marks_ignored_mask;
+	test_mask = event_mask & marks_mask & ~marks_ignore_mask;
 
 	/*
 	 * For dirent modification events (create/delete/move) that do not carry
@@ -411,7 +425,7 @@ static int fanotify_encode_fh(struct fanotify_fh *fh, struct inode *inode,
 	 * be zero in that case if encoding fh len failed.
 	 */
 	err = -ENOENT;
-	if (fh_len < 4 || WARN_ON_ONCE(fh_len % 4))
+	if (fh_len < 4 || WARN_ON_ONCE(fh_len % 4) || fh_len > MAX_HANDLE_SZ)
 		goto out_err;
 
 	/* No external buffer in a variable size allocated fh */
@@ -458,17 +472,41 @@ out_err:
 }
 
 /*
- * The inode to use as identifier when reporting fid depends on the event.
- * Report the modified directory inode on dirent modification events.
- * Report the "victim" inode otherwise.
+ * FAN_REPORT_FID is ambiguous in that it reports the fid of the child for
+ * some events and the fid of the parent for create/delete/move events.
+ *
+ * With the FAN_REPORT_TARGET_FID flag, the fid of the child is reported
+ * also in create/delete/move events in addition to the fid of the parent
+ * and the name of the child.
+ */
+static inline bool fanotify_report_child_fid(unsigned int fid_mode, u32 mask)
+{
+	if (mask & ALL_FSNOTIFY_DIRENT_EVENTS)
+		return (fid_mode & FAN_REPORT_TARGET_FID);
+
+	return (fid_mode & FAN_REPORT_FID) && !(mask & FAN_ONDIR);
+}
+
+/*
+ * The inode to use as identifier when reporting fid depends on the event
+ * and the group flags.
+ *
+ * With the group flag FAN_REPORT_TARGET_FID, always report the child fid.
+ *
+ * Without the group flag FAN_REPORT_TARGET_FID, report the modified directory
+ * fid on dirent events and the child fid otherwise.
+ *
  * For example:
- * 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.
+ * FS_ATTRIB reports the child fid even if reported on a watched parent.
+ * FS_CREATE reports the modified dir fid without FAN_REPORT_TARGET_FID.
+ *       and reports the created child fid with FAN_REPORT_TARGET_FID.
  */
 static struct inode *fanotify_fid_inode(u32 event_mask, const void *data,
-					int data_type, struct inode *dir)
+					int data_type, struct inode *dir,
+					unsigned int fid_mode)
 {
-	if (event_mask & ALL_FSNOTIFY_DIRENT_EVENTS)
+	if ((event_mask & ALL_FSNOTIFY_DIRENT_EVENTS) &&
+	    !(fid_mode & FAN_REPORT_TARGET_FID))
 		return dir;
 
 	return fsnotify_data_inode(data, data_type);
@@ -552,25 +590,34 @@ static struct fanotify_event *fanotify_alloc_fid_event(struct inode *id,
 	return &ffe->fae;
 }
 
-static struct fanotify_event *fanotify_alloc_name_event(struct inode *id,
+static struct fanotify_event *fanotify_alloc_name_event(struct inode *dir,
 							__kernel_fsid_t *fsid,
 							const struct qstr *name,
 							struct inode *child,
+							struct dentry *moved,
 							unsigned int *hash,
 							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);
+	struct inode *dir2 = moved ? d_inode(moved->d_parent) : NULL;
+	const struct qstr *name2 = moved ? &moved->d_name : NULL;
+	unsigned int dir_fh_len = fanotify_encode_fh_len(dir);
+	unsigned int dir2_fh_len = fanotify_encode_fh_len(dir2);
 	unsigned int child_fh_len = fanotify_encode_fh_len(child);
-	unsigned int size;
-
-	size = sizeof(*fne) + FANOTIFY_FH_HDR_LEN + dir_fh_len;
+	unsigned long name_len = name ? name->len : 0;
+	unsigned long name2_len = name2 ? name2->len : 0;
+	unsigned int len, size;
+
+	/* Reserve terminating null byte even for empty name */
+	size = sizeof(*fne) + name_len + name2_len + 2;
+	if (dir_fh_len)
+		size += FANOTIFY_FH_HDR_LEN + dir_fh_len;
+	if (dir2_fh_len)
+		size += FANOTIFY_FH_HDR_LEN + dir2_fh_len;
 	if (child_fh_len)
 		size += FANOTIFY_FH_HDR_LEN + child_fh_len;
-	if (name)
-		size += name->len + 1;
 	fne = kmalloc(size, gfp);
 	if (!fne)
 		return NULL;
@@ -580,24 +627,41 @@ static struct fanotify_event *fanotify_alloc_name_event(struct inode *id,
 	*hash ^= fanotify_hash_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, hash, 0);
+	if (dir_fh_len) {
+		dfh = fanotify_info_dir_fh(info);
+		len = fanotify_encode_fh(dfh, dir, dir_fh_len, hash, 0);
+		fanotify_info_set_dir_fh(info, len);
+	}
+	if (dir2_fh_len) {
+		dfh = fanotify_info_dir2_fh(info);
+		len = fanotify_encode_fh(dfh, dir2, dir2_fh_len, hash, 0);
+		fanotify_info_set_dir2_fh(info, len);
+	}
 	if (child_fh_len) {
 		ffh = fanotify_info_file_fh(info);
-		info->file_fh_totlen = fanotify_encode_fh(ffh, child,
-							child_fh_len, hash, 0);
+		len = fanotify_encode_fh(ffh, child, child_fh_len, hash, 0);
+		fanotify_info_set_file_fh(info, len);
 	}
-	if (name) {
-		long salt = name->len;
-
+	if (name_len) {
 		fanotify_info_copy_name(info, name);
-		*hash ^= full_name_hash((void *)salt, name->name, name->len);
+		*hash ^= full_name_hash((void *)name_len, name->name, name_len);
+	}
+	if (name2_len) {
+		fanotify_info_copy_name2(info, name2);
+		*hash ^= full_name_hash((void *)name2_len, name2->name,
+					name2_len);
 	}
 
-	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,
+	pr_debug("%s: size=%u dir_fh_len=%u child_fh_len=%u name_len=%u name='%.*s'\n",
+		 __func__, size, dir_fh_len, child_fh_len,
 		 info->name_len, info->name_len, fanotify_info_name(info));
 
+	if (dir2_fh_len) {
+		pr_debug("%s: dir2_fh_len=%u name2_len=%u name2='%.*s'\n",
+			 __func__, dir2_fh_len, info->name2_len,
+			 info->name2_len, fanotify_info_name2(info));
+	}
+
 	return &fne->fae;
 }
 
@@ -639,19 +703,21 @@ static struct fanotify_event *fanotify_alloc_error_event(
 	return &fee->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)
+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, u32 match_mask)
 {
 	struct fanotify_event *event = NULL;
 	gfp_t gfp = GFP_KERNEL_ACCOUNT;
-	struct inode *id = fanotify_fid_inode(mask, data, data_type, dir);
+	unsigned int fid_mode = FAN_GROUP_FLAG(group, FANOTIFY_FID_BITS);
+	struct inode *id = fanotify_fid_inode(mask, data, data_type, dir,
+					      fid_mode);
 	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 mem_cgroup *old_memcg;
+	struct dentry *moved = NULL;
 	struct inode *child = NULL;
 	bool name_event = false;
 	unsigned int hash = 0;
@@ -660,11 +726,10 @@ static struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
 
 	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
+		 * For certain events and group flags, report the child fid
 		 * in addition to reporting the parent fid and maybe child name.
 		 */
-		if ((fid_mode & FAN_REPORT_FID) && id != dirid && !ondir)
+		if (fanotify_report_child_fid(fid_mode, mask) && id != dirid)
 			child = id;
 
 		id = dirid;
@@ -688,6 +753,38 @@ static struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
 		} else if ((mask & ALL_FSNOTIFY_DIRENT_EVENTS) || !ondir) {
 			name_event = true;
 		}
+
+		/*
+		 * In the special case of FAN_RENAME event, use the match_mask
+		 * to determine if we need to report only the old parent+name,
+		 * only the new parent+name or both.
+		 * 'dirid' and 'file_name' are the old parent+name and
+		 * 'moved' has the new parent+name.
+		 */
+		if (mask & FAN_RENAME) {
+			bool report_old, report_new;
+
+			if (WARN_ON_ONCE(!match_mask))
+				return NULL;
+
+			/* Report both old and new parent+name if sb watching */
+			report_old = report_new =
+				match_mask & (1U << FSNOTIFY_ITER_TYPE_SB);
+			report_old |=
+				match_mask & (1U << FSNOTIFY_ITER_TYPE_INODE);
+			report_new |=
+				match_mask & (1U << FSNOTIFY_ITER_TYPE_INODE2);
+
+			if (!report_old) {
+				/* Do not report old parent+name */
+				dirid = NULL;
+				file_name = NULL;
+			}
+			if (report_new) {
+				/* Report new parent+name */
+				moved = fsnotify_data_dentry(data, data_type);
+			}
+		}
 	}
 
 	/*
@@ -709,9 +806,9 @@ static struct fanotify_event *fanotify_alloc_event(struct fsnotify_group *group,
 	} else if (fanotify_is_error_event(mask)) {
 		event = fanotify_alloc_error_event(group, fsid, data,
 						   data_type, &hash);
-	} else if (name_event && (file_name || child)) {
-		event = fanotify_alloc_name_event(id, fsid, file_name, child,
-						  &hash, gfp);
+	} else if (name_event && (file_name || moved || child)) {
+		event = fanotify_alloc_name_event(dirid, fsid, file_name, child,
+						  moved, &hash, gfp);
 	} else if (fid_mode) {
 		event = fanotify_alloc_fid_event(id, fsid, &hash, gfp);
 	} else {
@@ -743,16 +840,14 @@ out:
  */
 static __kernel_fsid_t fanotify_get_fsid(struct fsnotify_iter_info *iter_info)
 {
+	struct fsnotify_mark *mark;
 	int type;
 	__kernel_fsid_t fsid = {};
 
-	fsnotify_foreach_obj_type(type) {
+	fsnotify_foreach_iter_mark_type(iter_info, mark, type) {
 		struct fsnotify_mark_connector *conn;
 
-		if (!fsnotify_iter_should_report_type(iter_info, type))
-			continue;
-
-		conn = READ_ONCE(iter_info->marks[type]->connector);
+		conn = READ_ONCE(mark->connector);
 		/* Mark is just getting destroyed or created? */
 		if (!conn)
 			continue;
@@ -800,6 +895,7 @@ static int fanotify_handle_event(struct fsnotify_group *group, u32 mask,
 	struct fanotify_event *event;
 	struct fsnotify_event *fsn_event;
 	__kernel_fsid_t fsid = {};
+	u32 match_mask = 0;
 
 	BUILD_BUG_ON(FAN_ACCESS != FS_ACCESS);
 	BUILD_BUG_ON(FAN_MODIFY != FS_MODIFY);
@@ -821,15 +917,17 @@ static int fanotify_handle_event(struct fsnotify_group *group, u32 mask,
 	BUILD_BUG_ON(FAN_OPEN_EXEC != FS_OPEN_EXEC);
 	BUILD_BUG_ON(FAN_OPEN_EXEC_PERM != FS_OPEN_EXEC_PERM);
 	BUILD_BUG_ON(FAN_FS_ERROR != FS_ERROR);
+	BUILD_BUG_ON(FAN_RENAME != FS_RENAME);
 
-	BUILD_BUG_ON(HWEIGHT32(ALL_FANOTIFY_EVENT_BITS) != 20);
+	BUILD_BUG_ON(HWEIGHT32(ALL_FANOTIFY_EVENT_BITS) != 21);
 
-	mask = fanotify_group_event_mask(group, iter_info, mask, data,
-					 data_type, dir);
+	mask = fanotify_group_event_mask(group, iter_info, &match_mask,
+					 mask, data, data_type, dir);
 	if (!mask)
 		return 0;
 
-	pr_debug("%s: group=%p mask=%x\n", __func__, group, mask);
+	pr_debug("%s: group=%p mask=%x report_mask=%x\n", __func__,
+		 group, mask, match_mask);
 
 	if (fanotify_is_perm_event(mask)) {
 		/*
@@ -848,7 +946,7 @@ static int fanotify_handle_event(struct fsnotify_group *group, u32 mask,
 	}
 
 	event = fanotify_alloc_event(group, mask, data, data_type, dir,
-				     file_name, &fsid);
+				     file_name, &fsid, match_mask);
 	ret = -ENOMEM;
 	if (unlikely(!event)) {
 		/*