RFC: rename __USER_CAP* to __USER_PTR*?

[brooksdavis]

In upstream FreeBSD, I'm aiming to avoid referring to capabilities in generic contexts, preferring to talk about pointers. E.g., I'm not currently defining intcap_t and I've removed types like kuint64cap_t. As such, I wonder if __USER_CAP should become __USER_PTR (or something else, but basically something that doesn't refer to a capability.

[jrtc27]

That makes sense to me, and even in a hybrid context it makes sense, a user pointer is represented by a capability.

[brooksdavis]

I also wonder: we want to stick with shouty case or should we use __user_ptr*() or some unprefixed thing since these are a lot like PTRIN.

[bsdjhb]

USER_PTR works for me I think to match PTRIN.