Daniel Lustig

The cumulativity fix seems reasonable to me

Thanks @laokz. ACK the first change about the missing initial condition. NACK the second change. That text is attempting to explain an MMIO variant of message passing (MP). It's trying...

Yes, I agree about A.22, good catch.

The last example in Figure A.4 is covered by the two paragraphs that follow it: neither address mismatch nor intervening stores from the same thread technically force the SC to...

Thanks for the continued typo hunting. Figure A.4: yes, all of the `sc` instructions are missing an operand, thanks. However, the last example intentionally shows `8(s0)` rather than `0(s0)` to...

Looks good to me, aside from the one minor comment I just posted

> @wmaxey @daniellustig i think we can just merge this. Having this would simplify the docs of atomic_ref as well. 👍

LGTM

Added one minor comment, otherwise LGTM