[PATCH 4.9 056/120] btrfs: fix crash when tracepoint arguments are freed by wq callbacks

[Greg Kroah-Hartman]

4.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: David Sterba <dsterba@xxxxxxxx>

commit ac0c7cf8be00f269f82964cf7b144ca3edc5dbc4 upstream.

Enabling btrfs tracepoints leads to instant crash, as reported. The wq
callbacks could free the memory and the tracepoints started to
dereference the members to get to fs_info.

The proposed fix https://marc.info/?l=linux-btrfs&m=148172436722606&w=2
removed the tracepoints but we could preserve them by passing only the
required data in a safe way.

Fixes: bc074524e123 ("btrfs: prefix fsid to all trace events")
Reported-by: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx>
Reviewed-by: Qu Wenruo <quwenruo@xxxxxxxxxxxxxx>
Signed-off-by: David Sterba <dsterba@xxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 fs/btrfs/async-thread.c      |   15 +++++++++++----
 include/trace/events/btrfs.h |   22 +++++++++++++---------
 2 files changed, 24 insertions(+), 13 deletions(-)

--- a/fs/btrfs/async-thread.c
+++ b/fs/btrfs/async-thread.c
@@ -273,6 +273,8 @@ static void run_ordered_work(struct __bt
 	unsigned long flags;
 
 	while (1) {
+		void *wtag;
+
 		spin_lock_irqsave(lock, flags);
 		if (list_empty(list))
 			break;
@@ -299,11 +301,13 @@ static void run_ordered_work(struct __bt
 		spin_unlock_irqrestore(lock, flags);
 
 		/*
-		 * we don't want to call the ordered free functions
-		 * with the lock held though
+		 * We don't want to call the ordered free functions with the
+		 * lock held though. Save the work as tag for the trace event,
+		 * because the callback could free the structure.
 		 */
+		wtag = work;
 		work->ordered_free(work);
-		trace_btrfs_all_work_done(work);
+		trace_btrfs_all_work_done(wq->fs_info, wtag);
 	}
 	spin_unlock_irqrestore(lock, flags);
 }
@@ -311,6 +315,7 @@ static void run_ordered_work(struct __bt
 static void normal_work_helper(struct btrfs_work *work)
 {
 	struct __btrfs_workqueue *wq;
+	void *wtag;
 	int need_order = 0;
 
 	/*
@@ -324,6 +329,8 @@ static void normal_work_helper(struct bt
 	if (work->ordered_func)
 		need_order = 1;
 	wq = work->wq;
+	/* Safe for tracepoints in case work gets freed by the callback */
+	wtag = work;
 
 	trace_btrfs_work_sched(work);
 	thresh_exec_hook(wq);
@@ -333,7 +340,7 @@ static void normal_work_helper(struct bt
 		run_ordered_work(wq);
 	}
 	if (!need_order)
-		trace_btrfs_all_work_done(work);
+		trace_btrfs_all_work_done(wq->fs_info, wtag);
 }
 
 void btrfs_init_work(struct btrfs_work *work, btrfs_work_func_t uniq_func,
--- a/include/trace/events/btrfs.h
+++ b/include/trace/events/btrfs.h
@@ -1162,22 +1162,26 @@ DECLARE_EVENT_CLASS(btrfs__work,
 		   __entry->func, __entry->ordered_func, __entry->ordered_free)
 );
 
-/* For situiations that the work is freed */
+/*
+ * For situiations when the work is freed, we pass fs_info and a tag that that
+ * matches address of the work structure so it can be paired with the
+ * scheduling event.
+ */
 DECLARE_EVENT_CLASS(btrfs__work__done,
 
-	TP_PROTO(struct btrfs_work *work),
+	TP_PROTO(struct btrfs_fs_info *fs_info, void *wtag),
 
-	TP_ARGS(work),
+	TP_ARGS(fs_info, wtag),
 
 	TP_STRUCT__entry_btrfs(
-		__field(	void *,	work			)
+		__field(	void *,	wtag			)
 	),
 
-	TP_fast_assign_btrfs(btrfs_work_owner(work),
-		__entry->work		= work;
+	TP_fast_assign_btrfs(fs_info,
+		__entry->wtag		= wtag;
 	),
 
-	TP_printk_btrfs("work->%p", __entry->work)
+	TP_printk_btrfs("work->%p", __entry->wtag)
 );
 
 DEFINE_EVENT(btrfs__work, btrfs_work_queued,
@@ -1196,9 +1200,9 @@ DEFINE_EVENT(btrfs__work, btrfs_work_sch
 
 DEFINE_EVENT(btrfs__work__done, btrfs_all_work_done,
 
-	TP_PROTO(struct btrfs_work *work),
+	TP_PROTO(struct btrfs_fs_info *fs_info, void *wtag),
 
-	TP_ARGS(work)
+	TP_ARGS(fs_info, wtag)
 );
 
 DEFINE_EVENT(btrfs__work, btrfs_ordered_sched,