Search for a command to run...
Several existing dynamic binary analysis tools use shadowmemory-they shadow, in software, every byte of memory used by a program with another value that says something about it. Shadow memory is difficult to implement both efficiently and robustly. Nonetheless, existing shadow memory implementations have not been studied in detail. This is unfortunate, because shadow memory is powerful-for example, some of the existing tools that use it detect critical errors such as bad memory accesses, data races, and uses of uninitialised or untrusted data. In this paper we describe the implementation of shadow memory in Memcheck, a popular memory checker built with Valgrind, a dynamic binary instrumentation framework. This implementation has several novel features that make it efficient: carefully chosen data structures and operations result in a mean slow-down factor of only 22.2 and moderate memory usage. This may sound slow, but we show it is 8.9 times faster and 8.5 times smaller on average than a naive implementation, and shadow memory operations account for only about half of Memcheck's execution time. Equally importantly, unlike some tools, Memcheck's shadow memory implementation is robust: it is used on Linux by thousands of programmers on sizeable programs such as Mozilla and OpenOffice, and is suited to almost any memory configuration. This is the first detailed description of a robust shadow memory implementation, and the first detailed experimental evaluation of any shadow memory implementation. The ideas within are applicable to any shadow memory tool built with any instrumentation framework.