worker: ignore subsequent start messages

If the worker is already started, don't assert and just return.

This fixes calling spice_server_vm_start() multiple times.

Note: this was triggered as a race with the proposed QEMU patch: https://patchew.org/QEMU/20210127100206.277698-1-marcandre.lureau@redhat.com/, I am trying to solve it differently in QEMU now, but this patch still makes things a bit more robust,