Sculpt: Fix possible race condition with undo nodes

it is possible that two threads will request same undo node, only one
of them will initialize the node. The issue is that initialization is
happening outside of a lock, which was making one thread to use non-
initialized node.

If this change is ever a bottleneck, make a lock inside of node.

1 files changed, 6 insertions, 3 deletions

diff --git a/source/blender/editors/sculpt_paint/sculpt_undo.c b/source/blender/editors/sculpt_paint/sculpt_undo.c
index bc22147b15d..61f7c3bd0d9 100644
--- a/source/blender/editors/sculpt_paint/sculpt_undo.c
+++ b/source/blender/editors/sculpt_paint/sculpt_undo.c
@@ -931,9 +931,10 @@ SculptUndoNode *sculpt_undo_push_node(
 
 	unode = sculpt_undo_alloc_node(ob, node, type);
 
-	BLI_thread_unlock(LOCK_CUSTOM1);
-
-	/* copy threaded, hopefully this is the performance critical part */
+	/* NOTE: If this ever becomes a bottleneck, make a lock inside of the node.
+	 * so we release global lock sooner, but keep data locked for until it is
+	 * fully initialized.
+	 */
 
 	if (unode->grids) {
 		int totgrid, *grids;
@@ -970,6 +971,8 @@ SculptUndoNode *sculpt_undo_push_node(
 	if (ss->kb) BLI_strncpy(unode->shapeName, ss->kb->name, sizeof(ss->kb->name));
 	else unode->shapeName[0] = '\0';
 
+	BLI_thread_unlock(LOCK_CUSTOM1);
+
 	return unode;
 }