Patch: PKCS11-spy does not gracefully handle post-Finalization call attempts

[harningt]

Inside an application I am working on I found that it called Finalize twice (once explicitly, once in destructor).

In looking for a quick-fix (since the application is a bear to successfully compile in its current state) I found that nearly _all_ pkcs11-spy calls are unsafe. So I added a check for po being null before every call and if po == NULL, it returns CKR_CRYPTOKI_NOT_INITIALIZED since the only cause of po == NULL is non-initialization or early Finalization...