Alright, I need to classify this bug report regarding QEMU as either 'user-mode' or 'system-mode'. The issue is about undefined behavior in memcpy during write_erst_record. From what I know, QEMU has two main modes: user-mode and system-mode. User-mode emulation runs applications compiled for one architecture on another without needing the target OS to be present. It's more lightweight but limited in scope. System-mode is full-system emulation, where QEMU emulates an entire computer system, including the BIOS and OS, allowing it to run different operating systems as if they were on real hardware. The mention of 'memcpy' in a function related to writing records suggests this could be dealing with lower-level operations, possibly handling device or memory management. Undefined behavior here might indicate issues in how data is copied during system-level tasks, which are typically handled in system-mode where more complex and critical operations occur. Given that the problem involves higher privileges, peripherals, devices, and potentially the OS layer, it's likely a system-mode issue. system