* Revisit Boot Information Blob address passing for the different conditions of cold and warm boot for first and secondary execution contexts.
* Harden to SPMD to only accept a direct resp. from the SPMC after converying a 'power management' direct request
* Revisit cpu off and vCPU off usage in spmd handler:
* Create a vm helper function for use by plat_psci_cpu_resume and psci_primary_vm_handler