Fix post_singlestep for cases where event was already sunk
diff --git a/sysdeps/linux-gnu/trace.c b/sysdeps/linux-gnu/trace.c
index ef85747..149534f 100644
--- a/sysdeps/linux-gnu/trace.c
+++ b/sysdeps/linux-gnu/trace.c
@@ -617,7 +617,7 @@
 {
 	continue_for_sigstop_delivery(&self->pids);
 
-	if ((*eventp)->type == EVENT_BREAKPOINT)
+	if (*eventp != NULL && (*eventp)->type == EVENT_BREAKPOINT)
 		*eventp = NULL; // handled
 
 	if (self->atomic_skip_bp_addr != 0)