<p>Smart contracts are computer programs deployed on blockchains to facilitate transactions. A critical aspect of smart contract security is the use of state-reverting statements (e.g., <Emphasis FontCategory="NonProportional">require</Emphasis>, <Emphasis FontCategory="NonProportional">if...revert</Emphasis>, <Emphasis FontCategory="NonProportional">if...throw</Emphasis>). These statements protect transactions from abnormal behaviors or malicious attacks by reverting a contract to its previous state when certain input constraints or security properties are violated. While essential, the correct use of these state-reverting (SR) statements is nontrivial. Improper use can lead to security vulnerabilities, resulting in substantial financial losses or other severe consequences. It is, therefore, highly important to understand developers’ practices of state reverting in smart contracts and the common mistakes they make. To achieve this goal, we conduct the first comprehensive empirical study on the use of SR statements and their related faults in Solidity smart contracts. First, we analyze the prevalence and purposes of SR statements in 21,414 verified contracts from popular decentralized applications (dapps) and manually examine 381 SR statements, leading to a taxonomy of their uses. Second, we collect 320 real-world state-reverting faults (SR faults) from open-source projects on GitHub and audit reports on Code4rena. We categorize the SR faults into 17 types and summarize 12 distinct fixing strategies. This knowledge can help researchers and practitioners to better understand the common usages of SR statements and learn how to prevent or cope with SR faults. Lastly, the variety of SR fault types and the presence of high-risk issues highlight the need for automated tools to identify and mitigate these faults. This further motivates us to assess the SR fault detection performance of state-of-the-art security analyzers, with the aim of understanding their capability and identifying their deficiencies. Via evaluating 12 representative tools on a benchmark comprising 243 contracts with six types of SR faults and the corresponding patched versions, we observe that existing tools exhibit limited capabilities in detecting SR faults (the average detection rate is 14.4%). This result underscores the need for more advanced security analysis tools specifically tailored for SR faults. To facilitate the development of such tools, we further provide a comprehensive analysis of three common limitations of existing tools.</p>

On state reverting in solidity smart contracts: Developer practices, fault categorization, and tool evaluation

期刊论文
发表日期：

2025年7月17日

查看全文

Lu Liu,
Lili Wei,
Wuqi Zhang,
Shuqing Li,
Yifan Zhou,
Yepang Liu,
Shing-Chi Cheung,
Michael R. Lyu

摘要

Smart contracts are computer programs deployed on blockchains to facilitate transactions. A critical aspect of smart contract security is the use of state-reverting statements (e.g., require, if...revert, if...throw). These statements protect transactions from abnormal behaviors or malicious attacks by reverting a contract to its previous state when certain input constraints or security properties are violated. While essential, the correct use of these state-reverting (SR) statements is nontrivial. Improper use can lead to security vulnerabilities, resulting in substantial financial losses or other severe consequences. It is, therefore, highly important to understand developers’ practices of state reverting in smart contracts and the common mistakes they make. To achieve this goal, we conduct the first comprehensive empirical study on the use of SR statements and their related faults in Solidity smart contracts. First, we analyze the prevalence and purposes of SR statements in 21,414 verified contracts from popular decentralized applications (dapps) and manually examine 381 SR statements, leading to a taxonomy of their uses. Second, we collect 320 real-world state-reverting faults (SR faults) from open-source projects on GitHub and audit reports on Code4rena. We categorize the SR faults into 17 types and summarize 12 distinct fixing strategies. This knowledge can help researchers and practitioners to better understand the common usages of SR statements and learn how to prevent or cope with SR faults. Lastly, the variety of SR fault types and the presence of high-risk issues highlight the need for automated tools to identify and mitigate these faults. This further motivates us to assess the SR fault detection performance of state-of-the-art security analyzers, with the aim of understanding their capability and identifying their deficiencies. Via evaluating 12 representative tools on a benchmark comprising 243 contracts with six types of SR faults and the corresponding patched versions, we observe that existing tools exhibit limited capabilities in detecting SR faults (the average detection rate is 14.4%). This result underscores the need for more advanced security analysis tools specifically tailored for SR faults. To facilitate the development of such tools, we further provide a comprehensive analysis of three common limitations of existing tools.