Fix inconsistency in p/invoke arguments introduced when switching some to be blittable

[elinor-fung]

Resolves #61062

cc @AaronRobinsonMSFT @jkoritzinsky

[jkoritzinsky]

I think @stephentoub wanted us to go in the other (non-blittable byrefs) direction, not making everything pointers.

[AaronRobinsonMSFT]

LGTM.

[elinor-fung]

I had interpreted #59658 (comment) (and therefore #59658 (comment)) as making them all pointers, since in this case, unlike some of the bool<->int ones we had, it doesn't make them much less usable.

[jkoritzinsky]

I remember during our internal discussions that there was a dislike about losing the conceptual ideas that the out annotation brings (that the parameter doesn't need to be initialized before being passed to native code) that we may have forgotten to write down.

[elinor-fung]

Ah, I do recall discussion about out - for whatever reason, that I lumped that with the bool in my mind. Thoughts, @stephentoub?

[deeprobin]

LGTM

[elinor-fung]

Went back and looked at some notes and what they were originally - switched back to out/ref.