Linux Memory Forensics: Dissecting the User Space Process Heap

The analysis of memory during a forensic investigation is often an important step to reconstruct events. While prior work in this field has mostly concentrated on information residing in the kernel space (process lists, network connections, and so on) and in particular on the Microsoft Windows operating system, this work focuses on Linux user space processes as they might also contain valuable information for an investigation. Because a lot of process data is located in the heap, this work in the first place concentrates on the analysis of Glibc’s heap implementation and on how and where heap related information is stored in the virtual memory of Linux processes that use the Glibc heap implementation. Up to now, the heap was mostly considered a large cohesive memory region from a memory forensics perspective, making it rather hard manual work to identify relevant information inside. We introduce a Python class for the memory analysis framework Rekall that is based on our analysis results and allows access to all chunks contained in the heap and their meta information. Further, based on this class, six plugins have been developed that support an investigator in analyzing user space processes: Four of these plugins provide generic analysis capabilities such as finding information/references within chunks and dumping chunks into separate files for further investigation. These plugins have been used to reverse engineer data structures within the heap for user space processes, while illustrating how such plugins ease the whole analysis process. The remaining two plugins are a result of these user space process analyses and are extracting the command history for the zsh shell and password entry information for the password manager KeePassX. This report is an extended version of our paper published at DFRWS USA (Block and Dewald, 2017).