On Wed, 16 Nov 2022 at 08:19, Andres Freund <andres@anarazel.de> wrote:
> We already rely on memory context returning MAXIMUM_ALIGNOF aligned
> allocations. Adding the special case, I think, means that the we could safely
> over-allocate by "only"
> alignto + sizeof(MemoryChunk) - MAXIMUM_ALIGNOF
>
> Which would be a reasonable win for small allocations with a small >
> MAXIMUM_ALIGNOF alignment. But I don't think that'll be a very common case?
Seems reasonable. Subtracting MAXIMUM_ALIGNOF doesn't add any
additional run-time cost since it will be constant folded with the
sizeof(MemoryChunk).
I've attached an updated patch. The 0002 is just intended to exercise
these allocations a little bit, it's not intended for commit. I was
using that to ensure valgrind does not complain about anything. It
seems happy now.
David
