[foreign-memaccess+abi] RFR: 8299911: Refactor struct handling in AArch64 CallArranger [v3]

Wed Jan 11 13:53:37 UTC 2023

On Wed, 11 Jan 2023 13:41:31 GMT, Jorn Vernee <jvernee at openjdk.org> wrote:

>>> I've attempted to clarify the comment a bit.
>> 
>> Just to check - where do we check that these structs do not have any padding inside? Does a meaningful exception get produced if that is not the case?
>
> Only HFAs are handled in this field-wise manner (see the table in the comment on this method as well). The assumption is that any struct with padding doesn't get classified as an HFA (and they don't in the current code), and we can not get here (see also [`jdk.internal.foreign.abi.aarch64.TypeClass::isHomogeneousFloatAggregate`](https://github.com/openjdk/panama-foreign/blob/08225e4f5e00837e974098a933f27267f3c0dbc5/src/java.base/share/classes/jdk/internal/foreign/abi/aarch64/TypeClass.java#L61-L91)). But, since this is a non-local assumption, I added a comment as well.

I've attempted to clarify the comment once more.

-------------

PR: https://git.openjdk.org/panama-foreign/pull/765