Function profiling for embedded software by utilizing QEMU and analyzer tool
Tran Van Dung, Ittetsu Taniguchi, Takuji Hieda, Hiroyuki Tomiyama · 2013
Function profiling is crucial for optimized embedded software which needs to have resource constraint, low level power consumption and real-time ability. In this work, we provide a fast and reliable solution by utilizing an instruction set simulator named QEMU and creating an analyzer tool. We developed a tracing module inside the simulator to trace execution information of software in run-time and record it in a log file. Our implementation takes advantages of the dynamic binary translation of QEMU to keep its speed fast and use an analyzer tool to analyze the log file and creates a function profile. We implemented this methodology for ARM architecture, and evaluated many kinds of embedded applications.