Code review comment for lp:~alan-griffiths/mir/add-mir_surface_spec_attach

after chatting on IRC, seems the problem is wide enough to go with a packed-value or opcode approach for ABI stability. We have a packed-value approach in this MP, would slightly prefer an opcode approach, but am alright with the packed-value.

re MirPlacement discussion,
seeing as the future plans are to expand, MirPlacement seems to me to specific enough to suggest to the user what its for, while still being generic enough to allow different placement modes to be set, so I'm alright MirPlacement.

suppose that boils down to approve.