The other CPUs (not CPU#0) do not have a syscall trap set. The syscall handler for CPU#0 should panic.