[PATCH 4.9 1/3] fsnotify: Provide framework for dropping SRCU lock in ->handle_event

[Matthew Ruffell <matthew.ruffell@canonical.com>]

From: Jan Kara <jack@suse.cz>

commit abc77577a669f424c5d0c185b9994f2621c52aa4 upstream.

fanotify wants to drop fsnotify_mark_srcu lock when waiting for response
from userspace so that the whole notification subsystem is not blocked
during that time. This patch provides a framework for safely getting
mark reference for a mark found in the object list which pins the mark
in that list. We can then drop fsnotify_mark_srcu, wait for userspace
response and then safely continue iteration of the object list once we
reaquire fsnotify_mark_srcu.

Reviewed-by: Miklos Szeredi <mszeredi@redhat.com>
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Signed-off-by: Jan Kara <jack@suse.cz>
[mruffell: backport: realign file fs/notify/mark.c]
Signed-off-by: Matthew Ruffell <matthew.ruffell@canonical.com>
---
 fs/notify/fsnotify.h             |  6 +++
 fs/notify/group.c                |  1 +
 fs/notify/mark.c                 | 83 +++++++++++++++++++++++++++++++-
 include/linux/fsnotify_backend.h |  5 ++
 4 files changed, 94 insertions(+), 1 deletion(-)

diff --git a/fs/notify/fsnotify.h b/fs/notify/fsnotify.h
index 0a3bc2cf192c..0ad0eb9f2e14 100644
--- a/fs/notify/fsnotify.h
+++ b/fs/notify/fsnotify.h
@@ -8,6 +8,12 @@
 
 #include "../mount.h"
 
+struct fsnotify_iter_info {
+	struct fsnotify_mark *inode_mark;
+	struct fsnotify_mark *vfsmount_mark;
+	int srcu_idx;
+};
+
 /* destroy all events sitting in this groups notification queue */
 extern void fsnotify_flush_notify(struct fsnotify_group *group);
 
diff --git a/fs/notify/group.c b/fs/notify/group.c
index fbe3cbebec16..864103b707f4 100644
--- a/fs/notify/group.c
+++ b/fs/notify/group.c
@@ -124,6 +124,7 @@ struct fsnotify_group *fsnotify_alloc_group(const struct fsnotify_ops *ops)
 	/* set to 0 when there a no external references to this group */
 	atomic_set(&group->refcnt, 1);
 	atomic_set(&group->num_marks, 0);
+	atomic_set(&group->user_waits, 0);
 
 	spin_lock_init(&group->notification_lock);
 	INIT_LIST_HEAD(&group->notification_list);
diff --git a/fs/notify/mark.c b/fs/notify/mark.c
index d3fea0bd89e2..d3005d95d530 100644
--- a/fs/notify/mark.c
+++ b/fs/notify/mark.c
@@ -105,6 +105,16 @@ void fsnotify_get_mark(struct fsnotify_mark *mark)
 	atomic_inc(&mark->refcnt);
 }
 
+/*
+ * Get mark reference when we found the mark via lockless traversal of object
+ * list. Mark can be already removed from the list by now and on its way to be
+ * destroyed once SRCU period ends.
+ */
+static bool fsnotify_get_mark_safe(struct fsnotify_mark *mark)
+{
+	return atomic_inc_not_zero(&mark->refcnt);
+}
+
 void fsnotify_put_mark(struct fsnotify_mark *mark)
 {
 	if (atomic_dec_and_test(&mark->refcnt)) {
@@ -125,6 +135,72 @@ u32 fsnotify_recalc_mask(struct hlist_head *head)
 	return new_mask;
 }
 
+bool fsnotify_prepare_user_wait(struct fsnotify_iter_info *iter_info)
+{
+	struct fsnotify_group *group;
+
+	if (WARN_ON_ONCE(!iter_info->inode_mark && !iter_info->vfsmount_mark))
+		return false;
+
+	if (iter_info->inode_mark)
+		group = iter_info->inode_mark->group;
+	else
+		group = iter_info->vfsmount_mark->group;
+
+	/*
+	 * Since acquisition of mark reference is an atomic op as well, we can
+	 * be sure this inc is seen before any effect of refcount increment.
+	 */
+	atomic_inc(&group->user_waits);
+
+	if (iter_info->inode_mark) {
+		/* This can fail if mark is being removed */
+		if (!fsnotify_get_mark_safe(iter_info->inode_mark))
+			goto out_wait;
+	}
+	if (iter_info->vfsmount_mark) {
+		if (!fsnotify_get_mark_safe(iter_info->vfsmount_mark))
+			goto out_inode;
+	}
+
+	/*
+	 * Now that both marks are pinned by refcount in the inode / vfsmount
+	 * lists, we can drop SRCU lock, and safely resume the list iteration
+	 * once userspace returns.
+	 */
+	srcu_read_unlock(&fsnotify_mark_srcu, iter_info->srcu_idx);
+
+	return true;
+out_inode:
+	if (iter_info->inode_mark)
+		fsnotify_put_mark(iter_info->inode_mark);
+out_wait:
+	if (atomic_dec_and_test(&group->user_waits) && group->shutdown)
+		wake_up(&group->notification_waitq);
+	return false;
+}
+
+void fsnotify_finish_user_wait(struct fsnotify_iter_info *iter_info)
+{
+	struct fsnotify_group *group = NULL;
+
+	iter_info->srcu_idx = srcu_read_lock(&fsnotify_mark_srcu);
+	if (iter_info->inode_mark) {
+		group = iter_info->inode_mark->group;
+		fsnotify_put_mark(iter_info->inode_mark);
+	}
+	if (iter_info->vfsmount_mark) {
+		group = iter_info->vfsmount_mark->group;
+		fsnotify_put_mark(iter_info->vfsmount_mark);
+	}
+	/*
+	 * We abuse notification_waitq on group shutdown for waiting for all
+	 * marks pinned when waiting for userspace.
+	 */
+	if (atomic_dec_and_test(&group->user_waits) && group->shutdown)
+		wake_up(&group->notification_waitq);
+}
+
 /*
  * Remove mark from inode / vfsmount list, group list, drop inode reference
  * if we got one.
@@ -161,7 +237,6 @@ void fsnotify_detach_mark(struct fsnotify_mark *mark)
 	 * __fsnotify_parent() lazily when next event happens on one of our
 	 * children.
 	 */
-
 	list_del_init(&mark->g_list);
 
 	spin_unlock(&mark->lock);
@@ -508,6 +583,12 @@ void fsnotify_detach_group_marks(struct fsnotify_group *group)
 		__fsnotify_free_mark(mark);
 		fsnotify_put_mark(mark);
 	}
+	/*
+	 * Some marks can still be pinned when waiting for response from
+	 * userspace. Wait for those now. fsnotify_prepare_user_wait() will
+	 * not succeed now so this wait is race-free.
+	 */
+	wait_event(group->notification_waitq, !atomic_read(&group->user_waits));
 }
 
 void fsnotify_duplicate_mark(struct fsnotify_mark *new, struct fsnotify_mark *old)
diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h
index 79467b239fcf..3de53a2b8944 100644
--- a/include/linux/fsnotify_backend.h
+++ b/include/linux/fsnotify_backend.h
@@ -79,6 +79,7 @@ struct fsnotify_event;
 struct fsnotify_mark;
 struct fsnotify_event_private_data;
 struct fsnotify_fname;
+struct fsnotify_iter_info;
 
 /*
  * Each group much define these ops.  The fsnotify infrastructure will call
@@ -162,6 +163,8 @@ struct fsnotify_group {
 	struct fsnotify_event *overflow_event;	/* Event we queue when the
 						 * notification list is too
 						 * full */
+	atomic_t user_waits;		/* Number of tasks waiting for user
+					 * response */
 
 	/* groups can define private fields here or use the void *private */
 	union {
@@ -350,6 +353,8 @@ extern void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group, un
 extern void fsnotify_get_mark(struct fsnotify_mark *mark);
 extern void fsnotify_put_mark(struct fsnotify_mark *mark);
 extern void fsnotify_unmount_inodes(struct super_block *sb);
+extern void fsnotify_finish_user_wait(struct fsnotify_iter_info *iter_info);
+extern bool fsnotify_prepare_user_wait(struct fsnotify_iter_info *iter_info);
 
 /* put here because inotify does some weird stuff when destroying watches */
 extern void fsnotify_init_event(struct fsnotify_event *event,
-- 
2.19.1