Buffer overflow vulnerabilities in CUDA: a preliminary analysis
This addresses security risks for users of GPU-accelerated software, but it is incremental as it builds on known vulnerability analysis in a new context.
The paper investigates buffer overflow vulnerabilities in CUDA software on GPUs, demonstrating that attackers can exploit these to corrupt data or manipulate execution flow, such as by overwriting function pointers.
We present a preliminary study of buffer overflow vulnerabilities in CUDA software running on GPUs. We show how an attacker can overrun a buffer to corrupt sensitive data or steer the execution flow by overwriting function pointers, e.g., manipulating the virtual table of a C++ object. In view of a potential mass market diffusion of GPU accelerated software this may be a major concern.