Locked out by Latch-up? An Empirical Study on Laser Fault Injection into Arm Cortex-M Processors

Laser-based fault injection (LFI) is considered as one of the most powerful tools for active attacks against integrated circuits. However, only few empirical results are published for LFI into modern low-power microcontrollers with current process technologies. To fill this gap, we investigate LFI in four Cortex-M microcontrollers from different manufacturers: ST Microelectronics, NXP and Infineon. We note that those controllers differ from the ones used in high-security smartcard devices but argue that they are possibly built in similar process technologies making our results relevant for security evaluations. We were able to successfully inject precise faults into either the SRAM or the register file in all tested devices. We report our settings and fault maps in order to facilitate further fault attack investigations on these microcontrollers. As another contribution, we would like to emphasize the significant difficulties we encountered in some measurements due to the occurrence of latch-up effects. In many cases, the latch-up behavior of the integrated circuit prevented successful fault injections. This observation is largely underrepresented in scientific publications, which leads to an overestimation of the effectiveness of laser-based fault injection attacks under realistic circumstances.      «

Laser-based fault injection (LFI) is considered as one of the most powerful tools for active attacks against integrated circuits. However, only few empirical results are published for LFI into modern low-power microcontrollers with current process technologies. To fill this gap, we investigate LFI in four Cortex-M microcontrollers from different manufacturers: ST Microelectronics, NXP and Infineon. We note that those controllers differ from the ones used in high-security smartcard devices bu...     »