test interleaving load/store roundtrip

[folkertdev]

Well, turns out some of these were extremely broken on aarch64_be

[folkertdev]

cc @adamgemmell but really all this does is add more tests. The failures i encountered while making the changes indicate that this is all actually fairly well-tested.

[rustbot]

r? @Amanieu

rustbot has assigned @Amanieu.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: @Amanieu, @folkertdev, @sayantn
• @Amanieu, @folkertdev, @sayantn expanded to Amanieu, folkertdev, sayantn
• Random selection from Amanieu, sayantn

[folkertdev]

actually

r? sayantn

because we should probably merge this before #2025

[sayantn]

@folkertdev lgtm, but I am not too familiar with the semantics of ARMBE loads. This PR changes a lot of intrinsic implementations for ARMBE, and it would be nice if we could confirm that this is the correct behavior

[folkertdev]

Yeah, fair. I'm not too worried here because of the intrinsic tests that do end up exercising the functions that changed, but let's see.

[folkertdev]

Oh also note that I'm only actually touching the unsigned variants of these functions. They used to defer to the signed variants, but applied additional shuffling. Logically, that just does not make sense.

[sayantn]

That's a nice point, at least we are making it more consistent. I don't have a problem with merging this rn, as this will unblock some other PRs.