Accelerating Meta Data Checks for Software Correctness and Security

As high GHZ processors become prevalent, adding hardware support to ensure the correctness and security of programs will be just as important, for the average user, as further increases in processor performance. The goal of our research is to focus on developing compiler and hardware support for efficiently performing software checks that can be left on all of the time, even in production code releases, to provide a significant increase in the correctness and security of software. In this paper we focus on the performance of checking the correctness of pointers. We focus on pointers since a significant amount of bugs and security issues (buffer overflows) in programs are due to memory bugs resulting from incorrect usage of pointers. To determine if a pointer reference is correct many techniques require additional information to be kept track of called meta-data. The meta-data is checked when a pointer is dereferenced to verify some property about the pointer or the object. The first part of our paper focuses on where to efficiently keep track of this meta-data information and the overheads for performing safety checks like bounds checking and dangling pointer checks. We then focus on architecture extensions to reduce the overhead of these meta-data checks. We examine these optimizations in the presence of two meta-data checking applications – bounds checking and dangling pointer checks and show that we can reduce the overhead of these pointer checks from 148% down to 21% on average.