Merge tag 'fsnotify_for_v5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs

Pull fsnotify updates from Jan Kara:
 "fsnotify speedups when notification actually isn't used and support
  for identifying processes which caused fanotify events through pidfd
  instead of normal pid"

* tag 'fsnotify_for_v5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  fsnotify: optimize the case of no marks of any type
  fsnotify: count all objects with attached connectors
  fsnotify: count s_fsnotify_inode_refs for attached connectors
  fsnotify: replace igrab() with ihold() on attach connector
  fanotify: add pidfd support to the fanotify API
  fanotify: introduce a generic info record copying helper
  fanotify: minor cosmetic adjustments to fid labels
  kernel/pid.c: implement additional checks upon pidfd_create() parameters
  kernel/pid.c: remove static qualifier from pidfd_create()

fs/notify/fsnotify.c

@@ -87,15 +87,15 @@ static void fsnotify_unmount_inodes(struct super_block *sb)
 
 	if (iput_inode)
 		iput(iput_inode);
-	/* Wait for outstanding inode references from connectors */
-	wait_var_event(&sb->s_fsnotify_inode_refs,
-		       !atomic_long_read(&sb->s_fsnotify_inode_refs));
 }
 
 void fsnotify_sb_delete(struct super_block *sb)
 {
 	fsnotify_unmount_inodes(sb);
 	fsnotify_clear_marks_by_sb(sb);
+	/* Wait for outstanding object references from connectors */
+	wait_var_event(&sb->s_fsnotify_connectors,
+		       !atomic_long_read(&sb->s_fsnotify_connectors));
 }
 
 /*

fs/notify/fsnotify.h

@@ -27,6 +27,21 @@ static inline struct super_block *fsnotify_conn_sb(
 	return container_of(conn->obj, struct super_block, s_fsnotify_marks);
 }
 
+static inline struct super_block *fsnotify_connector_sb(
+				struct fsnotify_mark_connector *conn)
+{
+	switch (conn->type) {
+	case FSNOTIFY_OBJ_TYPE_INODE:
+		return fsnotify_conn_inode(conn)->i_sb;
+	case FSNOTIFY_OBJ_TYPE_VFSMOUNT:
+		return fsnotify_conn_mount(conn)->mnt.mnt_sb;
+	case FSNOTIFY_OBJ_TYPE_SB:
+		return fsnotify_conn_sb(conn);
+	default:
+		return NULL;
+	}
+}
+
 /* destroy all events sitting in this groups notification queue */
 extern void fsnotify_flush_notify(struct fsnotify_group *group);
 

fs/notify/mark.c

@@ -169,6 +169,37 @@ static void fsnotify_connector_destroy_workfn(struct work_struct *work)
 	}
 }
 
+static void fsnotify_get_inode_ref(struct inode *inode)
+{
+	ihold(inode);
+	atomic_long_inc(&inode->i_sb->s_fsnotify_connectors);
+}
+
+static void fsnotify_put_inode_ref(struct inode *inode)
+{
+	struct super_block *sb = inode->i_sb;
+
+	iput(inode);
+	if (atomic_long_dec_and_test(&sb->s_fsnotify_connectors))
+		wake_up_var(&sb->s_fsnotify_connectors);
+}
+
+static void fsnotify_get_sb_connectors(struct fsnotify_mark_connector *conn)
+{
+	struct super_block *sb = fsnotify_connector_sb(conn);
+
+	if (sb)
+		atomic_long_inc(&sb->s_fsnotify_connectors);
+}
+
+static void fsnotify_put_sb_connectors(struct fsnotify_mark_connector *conn)
+{
+	struct super_block *sb = fsnotify_connector_sb(conn);
+
+	if (sb && atomic_long_dec_and_test(&sb->s_fsnotify_connectors))
+		wake_up_var(&sb->s_fsnotify_connectors);
+}
+
 static void *fsnotify_detach_connector_from_object(
 					struct fsnotify_mark_connector *conn,
 					unsigned int *type)
@@ -182,13 +213,13 @@ static void *fsnotify_detach_connector_from_object(
 	if (conn->type == FSNOTIFY_OBJ_TYPE_INODE) {
 		inode = fsnotify_conn_inode(conn);
 		inode->i_fsnotify_mask = 0;
-		atomic_long_inc(&inode->i_sb->s_fsnotify_inode_refs);
 	} else if (conn->type == FSNOTIFY_OBJ_TYPE_VFSMOUNT) {
 		fsnotify_conn_mount(conn)->mnt_fsnotify_mask = 0;
 	} else if (conn->type == FSNOTIFY_OBJ_TYPE_SB) {
 		fsnotify_conn_sb(conn)->s_fsnotify_mask = 0;
 	}
 
+	fsnotify_put_sb_connectors(conn);
 	rcu_assign_pointer(*(conn->obj), NULL);
 	conn->obj = NULL;
 	conn->type = FSNOTIFY_OBJ_TYPE_DETACHED;
@@ -209,19 +240,12 @@ static void fsnotify_final_mark_destroy(struct fsnotify_mark *mark)
 /* Drop object reference originally held by a connector */
 static void fsnotify_drop_object(unsigned int type, void *objp)
 {
-	struct inode *inode;
-	struct super_block *sb;
-
 	if (!objp)
 		return;
 	/* Currently only inode references are passed to be dropped */
 	if (WARN_ON_ONCE(type != FSNOTIFY_OBJ_TYPE_INODE))
 		return;
-	inode = objp;
-	sb = inode->i_sb;
-	iput(inode);
-	if (atomic_long_dec_and_test(&sb->s_fsnotify_inode_refs))
-		wake_up_var(&sb->s_fsnotify_inode_refs);
+	fsnotify_put_inode_ref(objp);
 }
 
 void fsnotify_put_mark(struct fsnotify_mark *mark)
@@ -493,8 +517,12 @@ static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
 		conn->fsid.val[0] = conn->fsid.val[1] = 0;
 		conn->flags = 0;
 	}
-	if (conn->type == FSNOTIFY_OBJ_TYPE_INODE)
-		inode = igrab(fsnotify_conn_inode(conn));
+	if (conn->type == FSNOTIFY_OBJ_TYPE_INODE) {
+		inode = fsnotify_conn_inode(conn);
+		fsnotify_get_inode_ref(inode);
+	}
+	fsnotify_get_sb_connectors(conn);
+
 	/*
 	 * cmpxchg() provides the barrier so that readers of *connp can see
 	 * only initialized structure
@@ -502,7 +530,7 @@ static int fsnotify_attach_connector_to_object(fsnotify_connp_t *connp,
 	if (cmpxchg(connp, NULL, conn)) {
 		/* Someone else created list structure for us */
 		if (inode)
-			iput(inode);
+			fsnotify_put_inode_ref(inode);
 		kmem_cache_free(fsnotify_mark_connector_cachep, conn);
 	}
 

include/linux/fanotify.h

@@ -27,6 +27,8 @@ extern struct ctl_table fanotify_table[]; /* for sysctl */
 
 #define FANOTIFY_FID_BITS	(FAN_REPORT_FID | FAN_REPORT_DFID_NAME)
 
+#define FANOTIFY_INFO_MODES	(FANOTIFY_FID_BITS | FAN_REPORT_PIDFD)
+
 /*
  * fanotify_init() flags that require CAP_SYS_ADMIN.
  * We do not allow unprivileged groups to request permission events.
@@ -35,6 +37,7 @@ extern struct ctl_table fanotify_table[]; /* for sysctl */
  */
 #define FANOTIFY_ADMIN_INIT_FLAGS	(FANOTIFY_PERM_CLASSES | \
 					 FAN_REPORT_TID | \
+					 FAN_REPORT_PIDFD | \
 					 FAN_UNLIMITED_QUEUE | \
 					 FAN_UNLIMITED_MARKS)
 

include/linux/fs.h

@@ -1507,8 +1507,11 @@ struct super_block {
 	/* Number of inodes with nlink == 0 but still referenced */
 	atomic_long_t s_remove_count;
 
-	/* Pending fsnotify inode refs */
-	atomic_long_t s_fsnotify_inode_refs;
+	/*
+	 * Number of inode/mount/sb objects that are being watched, note that
+	 * inodes objects are currently double-accounted.
+	 */
+	atomic_long_t s_fsnotify_connectors;
 
 	/* Being remounted read-only */
 	int s_readonly_remount;

include/linux/fsnotify.h

@@ -30,6 +30,9 @@ static inline void fsnotify_name(struct inode *dir, __u32 mask,
 				 struct inode *child,
 				 const struct qstr *name, u32 cookie)
 {
+	if (atomic_long_read(&dir->i_sb->s_fsnotify_connectors) == 0)
+		return;
+
 	fsnotify(mask, child, FSNOTIFY_EVENT_INODE, dir, name, NULL, cookie);
 }
 
@@ -41,6 +44,9 @@ static inline void fsnotify_dirent(struct inode *dir, struct dentry *dentry,
 
 static inline void fsnotify_inode(struct inode *inode, __u32 mask)
 {
+	if (atomic_long_read(&inode->i_sb->s_fsnotify_connectors) == 0)
+		return;
+
 	if (S_ISDIR(inode->i_mode))
 		mask |= FS_ISDIR;
 
@@ -53,6 +59,9 @@ static inline int fsnotify_parent(struct dentry *dentry, __u32 mask,
 {
 	struct inode *inode = d_inode(dentry);
 
+	if (atomic_long_read(&inode->i_sb->s_fsnotify_connectors) == 0)
+		return 0;
+
 	if (S_ISDIR(inode->i_mode)) {
 		mask |= FS_ISDIR;
 

include/linux/pid.h

@@ -78,6 +78,7 @@ struct file;
 
 extern struct pid *pidfd_pid(const struct file *file);
 struct pid *pidfd_get_pid(unsigned int fd, unsigned int *flags);
+int pidfd_create(struct pid *pid, unsigned int flags);
 
 static inline struct pid *get_pid(struct pid *pid)
 {

include/uapi/linux/fanotify.h

@@ -51,6 +51,7 @@
 #define FAN_ENABLE_AUDIT	0x00000040
 
 /* Flags to determine fanotify event format */
+#define FAN_REPORT_PIDFD	0x00000080	/* Report pidfd for event->pid */
 #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 */
@@ -123,6 +124,7 @@ 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
+#define FAN_EVENT_INFO_TYPE_PIDFD	4
 
 /* Variable length info record following event metadata */
 struct fanotify_event_info_header {
@@ -148,6 +150,15 @@ struct fanotify_event_info_fid {
 	unsigned char handle[0];
 };
 
+/*
+ * This structure is used for info records of type FAN_EVENT_INFO_TYPE_PIDFD.
+ * It holds a pidfd for the pid that was responsible for generating an event.
+ */
+struct fanotify_event_info_pidfd {
+	struct fanotify_event_info_header hdr;
+	__s32 pidfd;
+};
+
 struct fanotify_response {
 	__s32 fd;
 	__u32 response;
@@ -160,6 +171,8 @@ struct fanotify_response {
 
 /* No fd set in event */
 #define FAN_NOFD	-1
+#define FAN_NOPIDFD	FAN_NOFD
+#define FAN_EPIDFD	-2
 
 /* Helper functions to deal with fanotify_event_metadata buffers */
 #define FAN_EVENT_METADATA_LEN (sizeof(struct fanotify_event_metadata))

kernel/pid.c

@@ -550,13 +550,21 @@ struct pid *pidfd_get_pid(unsigned int fd, unsigned int *flags)
  * Note, that this function can only be called after the fd table has
  * been unshared to avoid leaking the pidfd to the new process.
  *
+ * This symbol should not be explicitly exported to loadable modules.
+ *
  * Return: On success, a cloexec pidfd is returned.
  *         On error, a negative errno number will be returned.
  */
-static int pidfd_create(struct pid *pid, unsigned int flags)
+int pidfd_create(struct pid *pid, unsigned int flags)
 {
 	int fd;
 
+	if (!pid || !pid_has_task(pid, PIDTYPE_TGID))
+		return -EINVAL;
+
+	if (flags & ~(O_NONBLOCK | O_RDWR | O_CLOEXEC))
+		return -EINVAL;
+
 	fd = anon_inode_getfd("[pidfd]", &pidfd_fops, get_pid(pid),
 			      flags | O_RDWR | O_CLOEXEC);
 	if (fd < 0)
@@ -596,10 +604,7 @@ SYSCALL_DEFINE2(pidfd_open, pid_t, pid, unsigned int, flags)
 	if (!p)
 		return -ESRCH;
 
-	if (pid_has_task(p, PIDTYPE_TGID))
-		fd = pidfd_create(p, flags);
-	else
-		fd = -EINVAL;
+	fd = pidfd_create(p, flags);
 
 	put_pid(p);
 	return fd;