simulation: properly setup SMP on qemu

We always use the EL2 qemu implementation, this allows the
elfloader to boot secondary cores using SMC (the only method it
supports). Note this is fine even if CONFIG_ARM_HYPERVISOR_SUPPORT is
not set as the elfloader will drop from EL2 to EL1 before entering the
seL4 kernel.

We also set the -smp flag to KernelMaxNumNodes, to launch the
necessary CPUs.

cmake-tool/helpers/simulation.cmake

@@ -144,14 +144,19 @@ function(GenerateSimulateScript)
         set(qemu_sim_extra_args "-bios none")
     elseif(KernelPlatformQEMUArmVirt)
         set(QemuBinaryMachine "qemu-system-${QEMU_ARCH}")
-        if(KernelArmHypervisorSupport)
+        if(Kernel32)
             set(sim_machine "virt,virtualization=on,highmem=off,secure=off")
         else()
-            set(sim_machine "virt")
+            set(sim_machine "virt,virtualization=on,highmem=on,secure=off")
         endif()
         set(sim_graphic_opt "-nographic")
         set(sim_cpu "${KernelArmCPU}")
         SetDefaultMemSize("${QEMU_MEMORY}")
+        if(KernelMaxNumNodes)
+            set(qemu_sim_extra_args "-smp ${KernelMaxNumNodes}")
+        else()
+            set(qemu_sim_extra_args "-smp 1")
+        endif()
     elseif(KernelPlatformQEMURiscVVirt)
         set(QemuBinaryMachine "qemu-system-${KernelSel4Arch}")
         set(sim_machine "virt")