Merge tag 'fsnotify_for_v5.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull fsnotify updates from Jan Kara:
 "The biggest part of this is support for fsnotify inode marks that
  don't pin inodes in memory but rather get evicted together with the
  inode (they are useful if userspace needs to exclude receipt of events
  from potentially large subtrees using fanotify ignore marks).

  There is also a fix for more consistent handling of events sent to
  parent and a fix of sparse(1) complaints"

* tag 'fsnotify_for_v5.19-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  fanotify: fix incorrect fmode_t casts
  fsnotify: consistent behavior for parent not watching children
  fsnotify: introduce mark type iterator
  fanotify: enable "evictable" inode marks
  fanotify: use fsnotify group lock helpers
  fanotify: implement "evictable" inode marks
  fanotify: factor out helper fanotify_mark_update_flags()
  fanotify: create helper fanotify_mark_user_flags()
  fsnotify: allow adding an inode mark without pinning inode
  dnotify: use fsnotify group lock helpers
  nfsd: use fsnotify group lock helpers
  audit: use fsnotify group lock helpers
  inotify: use fsnotify group lock helpers
  fsnotify: create helpers for group mark_mutex lock
  fsnotify: make allow_dups a property of the group
  fsnotify: pass flags argument to fsnotify_alloc_group()
  fsnotify: fix wrong lockdep annotations
  inotify: move control flags from mask to mark flags
  inotify: show inotify mask flags in proc fdinfo

3 files changed, 78 insertions, 22 deletions

diff --git a/include/linux/fanotify.h b/include/linux/fanotify.h
index 419cadcd7ff5..edc28555814c 100644
--- a/include/linux/fanotify.h
+++ b/include/linux/fanotify.h
@@ -66,6 +66,7 @@
 				 FAN_MARK_ONLYDIR | \
 				 FAN_MARK_IGNORED_MASK | \
 				 FAN_MARK_IGNORED_SURV_MODIFY | \
+				 FAN_MARK_EVICTABLE | \
 				 FAN_MARK_FLUSH)
 
 /*
diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h
index 0805b74cae44..9560734759fa 100644
--- a/include/linux/fsnotify_backend.h
+++ b/include/linux/fsnotify_backend.h
@@ -20,6 +20,7 @@
 #include <linux/user_namespace.h>
 #include <linux/refcount.h>
 #include <linux/mempool.h>
+#include <linux/sched/mm.h>
 
 /*
  * IN_* from inotfy.h lines up EXACTLY with FS_*, this is so we can easily
@@ -55,7 +56,6 @@
 #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_EXCL_UNLINK		0x04000000	/* do not send events if object is unlinked */
 /*
  * Set on inode mark that cares about things that happen to its children.
  * Always set for dnotify and inotify.
@@ -66,7 +66,6 @@
 #define FS_RENAME		0x10000000	/* File was renamed */
 #define FS_DN_MULTISHOT		0x20000000	/* dnotify multishot */
 #define FS_ISDIR		0x40000000	/* event occurred against dir */
-#define FS_IN_ONESHOT		0x80000000	/* only send event once */
 
 #define FS_MOVE			(FS_MOVED_FROM | FS_MOVED_TO)
 
@@ -106,8 +105,7 @@
 			     FS_ERROR)
 
 /* Extra flags that may be reported with event or control handling of events */
-#define ALL_FSNOTIFY_FLAGS  (FS_EXCL_UNLINK | FS_ISDIR | FS_IN_ONESHOT | \
-			     FS_DN_MULTISHOT | FS_EVENT_ON_CHILD)
+#define ALL_FSNOTIFY_FLAGS  (FS_ISDIR | FS_EVENT_ON_CHILD | FS_DN_MULTISHOT)
 
 #define ALL_FSNOTIFY_BITS   (ALL_FSNOTIFY_EVENTS | ALL_FSNOTIFY_FLAGS)
 
@@ -213,6 +211,12 @@ struct fsnotify_group {
 	unsigned int priority;
 	bool shutdown;		/* group is being shut down, don't queue more events */
 
+#define FSNOTIFY_GROUP_USER	0x01 /* user allocated group */
+#define FSNOTIFY_GROUP_DUPS	0x02 /* allow multiple marks per object */
+#define FSNOTIFY_GROUP_NOFS	0x04 /* group lock is not direct reclaim safe */
+	int flags;
+	unsigned int owner_flags;	/* stored flags of mark_mutex owner */
+
 	/* stores all fastpath marks assoc with this group so they can be cleaned on unregister */
 	struct mutex mark_mutex;	/* protect marks_list */
 	atomic_t user_waits;		/* Number of tasks waiting for user
@@ -253,6 +257,31 @@ struct fsnotify_group {
 	};
 };
 
+/*
+ * These helpers are used to prevent deadlock when reclaiming inodes with
+ * evictable marks of the same group that is allocating a new mark.
+ */
+static inline void fsnotify_group_lock(struct fsnotify_group *group)
+{
+	mutex_lock(&group->mark_mutex);
+	if (group->flags & FSNOTIFY_GROUP_NOFS)
+		group->owner_flags = memalloc_nofs_save();
+}
+
+static inline void fsnotify_group_unlock(struct fsnotify_group *group)
+{
+	if (group->flags & FSNOTIFY_GROUP_NOFS)
+		memalloc_nofs_restore(group->owner_flags);
+	mutex_unlock(&group->mark_mutex);
+}
+
+static inline void fsnotify_group_assert_locked(struct fsnotify_group *group)
+{
+	WARN_ON_ONCE(!mutex_is_locked(&group->mark_mutex));
+	if (group->flags & FSNOTIFY_GROUP_NOFS)
+		WARN_ON_ONCE(!(current->flags & PF_MEMALLOC_NOFS));
+}
+
 /* When calling fsnotify tell it if the data is a path or inode */
 enum fsnotify_data_type {
 	FSNOTIFY_EVENT_NONE,
@@ -370,6 +399,7 @@ static inline bool fsnotify_valid_obj_type(unsigned int obj_type)
 
 struct fsnotify_iter_info {
 	struct fsnotify_mark *marks[FSNOTIFY_ITER_TYPE_COUNT];
+	struct fsnotify_group *current_group;
 	unsigned int report_mask;
 	int srcu_idx;
 };
@@ -386,20 +416,31 @@ static inline void fsnotify_iter_set_report_type(
 	iter_info->report_mask |= (1U << iter_type);
 }
 
-static inline void fsnotify_iter_set_report_type_mark(
-		struct fsnotify_iter_info *iter_info, int iter_type,
-		struct fsnotify_mark *mark)
+static inline struct fsnotify_mark *fsnotify_iter_mark(
+		struct fsnotify_iter_info *iter_info, int iter_type)
 {
-	iter_info->marks[iter_type] = mark;
-	iter_info->report_mask |= (1U << iter_type);
+	if (fsnotify_iter_should_report_type(iter_info, iter_type))
+		return iter_info->marks[iter_type];
+	return NULL;
+}
+
+static inline int fsnotify_iter_step(struct fsnotify_iter_info *iter, int type,
+				     struct fsnotify_mark **markp)
+{
+	while (type < FSNOTIFY_ITER_TYPE_COUNT) {
+		*markp = fsnotify_iter_mark(iter, type);
+		if (*markp)
+			break;
+		type++;
+	}
+	return type;
 }
 
 #define FSNOTIFY_ITER_FUNCS(name, NAME) \
 static inline struct fsnotify_mark *fsnotify_iter_##name##_mark( \
 		struct fsnotify_iter_info *iter_info) \
 { \
-	return (iter_info->report_mask & (1U << FSNOTIFY_ITER_TYPE_##NAME)) ? \
-		iter_info->marks[FSNOTIFY_ITER_TYPE_##NAME] : NULL; \
+	return fsnotify_iter_mark(iter_info, FSNOTIFY_ITER_TYPE_##NAME); \
 }
 
 FSNOTIFY_ITER_FUNCS(inode, INODE)
@@ -409,6 +450,11 @@ FSNOTIFY_ITER_FUNCS(sb, SB)
 
 #define fsnotify_foreach_iter_type(type) \
 	for (type = 0; type < FSNOTIFY_ITER_TYPE_COUNT; type++)
+#define fsnotify_foreach_iter_mark_type(iter, mark, type) \
+	for (type = 0; \
+	     type = fsnotify_iter_step(iter, type, &mark), \
+	     type < FSNOTIFY_ITER_TYPE_COUNT; \
+	     type++)
 
 /*
  * fsnotify_connp_t is what we embed in objects which connector can be attached
@@ -427,6 +473,7 @@ struct fsnotify_mark_connector {
 	spinlock_t lock;
 	unsigned short type;	/* Type of object [lock] */
 #define FSNOTIFY_CONN_FLAG_HAS_FSID	0x01
+#define FSNOTIFY_CONN_FLAG_HAS_IREF	0x02
 	unsigned short flags;	/* flags [lock] */
 	__kernel_fsid_t fsid;	/* fsid of filesystem containing object */
 	union {
@@ -473,9 +520,15 @@ struct fsnotify_mark {
 	struct fsnotify_mark_connector *connector;
 	/* Events types to ignore [mark->lock, group->mark_mutex] */
 	__u32 ignored_mask;
-#define FSNOTIFY_MARK_FLAG_IGNORED_SURV_MODIFY	0x01
-#define FSNOTIFY_MARK_FLAG_ALIVE		0x02
-#define FSNOTIFY_MARK_FLAG_ATTACHED		0x04
+	/* General fsnotify mark flags */
+#define FSNOTIFY_MARK_FLAG_ALIVE		0x0001
+#define FSNOTIFY_MARK_FLAG_ATTACHED		0x0002
+	/* inotify mark flags */
+#define FSNOTIFY_MARK_FLAG_EXCL_UNLINK		0x0010
+#define FSNOTIFY_MARK_FLAG_IN_ONESHOT		0x0020
+	/* fanotify mark flags */
+#define FSNOTIFY_MARK_FLAG_IGNORED_SURV_MODIFY	0x0100
+#define FSNOTIFY_MARK_FLAG_NO_IREF		0x0200
 	unsigned int flags;		/* flags [mark->lock] */
 };
 
@@ -541,8 +594,9 @@ static inline void fsnotify_update_flags(struct dentry *dentry)
 /* called from fsnotify listeners, such as fanotify or dnotify */
 
 /* create a new group */
-extern struct fsnotify_group *fsnotify_alloc_group(const struct fsnotify_ops *ops);
-extern struct fsnotify_group *fsnotify_alloc_user_group(const struct fsnotify_ops *ops);
+extern struct fsnotify_group *fsnotify_alloc_group(
+				const struct fsnotify_ops *ops,
+				int flags);
 /* get reference to a group */
 extern void fsnotify_get_group(struct fsnotify_group *group);
 /* drop reference on a group from fsnotify_alloc_group */
@@ -635,26 +689,26 @@ extern int fsnotify_get_conn_fsid(const struct fsnotify_mark_connector *conn,
 /* attach the mark to the object */
 extern int fsnotify_add_mark(struct fsnotify_mark *mark,
 			     fsnotify_connp_t *connp, unsigned int obj_type,
-			     int allow_dups, __kernel_fsid_t *fsid);
+			     int add_flags, __kernel_fsid_t *fsid);
 extern int fsnotify_add_mark_locked(struct fsnotify_mark *mark,
 				    fsnotify_connp_t *connp,
-				    unsigned int obj_type, int allow_dups,
+				    unsigned int obj_type, int add_flags,
 				    __kernel_fsid_t *fsid);
 
 /* attach the mark to the inode */
 static inline int fsnotify_add_inode_mark(struct fsnotify_mark *mark,
 					  struct inode *inode,
-					  int allow_dups)
+					  int add_flags)
 {
 	return fsnotify_add_mark(mark, &inode->i_fsnotify_marks,
-				 FSNOTIFY_OBJ_TYPE_INODE, allow_dups, NULL);
+				 FSNOTIFY_OBJ_TYPE_INODE, add_flags, NULL);
 }
 static inline int fsnotify_add_inode_mark_locked(struct fsnotify_mark *mark,
 						 struct inode *inode,
-						 int allow_dups)
+						 int add_flags)
 {
 	return fsnotify_add_mark_locked(mark, &inode->i_fsnotify_marks,
-					FSNOTIFY_OBJ_TYPE_INODE, allow_dups,
+					FSNOTIFY_OBJ_TYPE_INODE, add_flags,
 					NULL);
 }
 
diff --git a/include/uapi/linux/fanotify.h b/include/uapi/linux/fanotify.h
index e8ac38cc2fd6..f1f89132d60e 100644
--- a/include/uapi/linux/fanotify.h
+++ b/include/uapi/linux/fanotify.h
@@ -82,6 +82,7 @@
 #define FAN_MARK_IGNORED_SURV_MODIFY	0x00000040
 #define FAN_MARK_FLUSH		0x00000080
 /* FAN_MARK_FILESYSTEM is	0x00000100 */
+#define FAN_MARK_EVICTABLE	0x00000200
 
 /* These are NOT bitwise flags.  Both bits can be used togther.  */
 #define FAN_MARK_INODE		0x00000000