On 2017-08-16 14:09:08 -0400, Tom Lane wrote:
> >> I'm not sure that that's good enough, and I'm damn sure that it
> >> shouldn't be undocumented.
>
> > 8 byte alignment would be good enough, so BUFFERALIGN ought to be
> > sufficient. But it'd be nicer to have a separate more descriptive knob.
>
> What I meant by possibly not good enough is that pg_atomic_uint64 used
> in other places isn't going to be very safe.
Well, it's not used otherwise in core so far, leaving test code
aside. It's correctly aligned if part of a aligned struct - the atomics
code itself can't really do anything about aligning that struct itself
isn't aligned.
> We might be effectively all right as long as we have a coding rule that
> pg_atomic_uint64 can only be placed in memory handed out by ShmemAlloc
> or shm_toc_allocate, which both have bigger-than-MAXALIGN alignment
> practices. But this needs to be documented.
Well, one could argue the alignment checks in every function are that
:). But yea, we probably should mention it more than that.
Greetings,
Andres Freund