Abstract
With the development of open source software and open source community, there are more available codes on the Internet. And the open vulnerability information can be found on the Internet. In fact, using known vulnerabilities to calculate the similarity with the source code has been demonstrated a useful method to detect vulnerabilities. But the vulnerabilities often have many irrelevant codes, which may cause false positives and reduce the accuracy of vulnerability detection. Besides, the program code may have been patched. This also leads to false positives. We use code property graphs to extract source code and calculate the similarity between the vulnerable code and the source code to judge whether the software has vulnerabilities. By using the patched code, we can reduce the false positive. We use our approach on LibTIFF and Linux kernel. The experimental results show that the approach can effectively find vulnerabilities and reduce the false positive.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have