Andres Freund <andres@anarazel.de> writes:
> On 2016-04-14 13:11:43 -0400, Tom Lane wrote:
>> But looking around a bit, I think that ReorderBufferRestoreChange is
>> almost completely broken on alignment-picky machines, if the expectation
>> is that the input data could be aligned arbitrarily.
> 'data' cannot be aligned arbitrarily. It's palloc aligned, and the
> buffer is used separately for everything is loaded from disk. I.e. the
> data pointer passed to ReorderBufferRestoreChange is aligned to a
> MAXALIGN boundary.
Right, I figured that out awhile later; it would have been better if it
was documented. See my second patch attempt.
regards, tom lane