Yunlong Xing, Shu Wang, Shiyu Sun, Xu He, and Kun Sun, George Mason University; Qi Li, Tsinghua University
Null pointer dereference (NPD) errors pose the risk of unexpected behavior and system instability, potentially leading to abrupt program termination due to exceptions or segmentation faults. When generating NPD fixes, all existing solutions are confined to the function level fixes and ignore the valuable intraprocedural and interprocedural contextual information, potentially resulting in incorrect patches. In this paper, we introduce CONCH, a novel approach that addresses the challenges of generating correct fixes for NPD issues by incorporating contextual checks. Our method first constructs an NPD context graph to maintain the semantics related to patch generation. Then we summarize distinct fixing position selection policies based on the distribution of the error positions, ensuring the resolution of bugs without introducing duplicate code. Next, the intraprocedural state retrogression builds the if condition, retrogresses the local resources, and constructs return statements as an initial patch. Finally, we conduct interprocedural state propagation to assess the correctness of the initial patch in the entire call chain. We evaluate the effectiveness of CONCH over two real-world datasets. The experimental results demonstrate that CONCH outperforms the SOTA methods and yields over 85% accurate patches.
Open Access Media
USENIX is committed to Open Access to the research presented at our events. Papers and proceedings are freely available to everyone once the event begins. Any video, audio, and/or slides that are posted after the event are also free and open to everyone. Support USENIX and our commitment to Open Access.