RFC: do we want to upstream vm_pointer_t?

[brooksdavis]

IIRC @qwattash chose vm_pointer_t as an easy way to keep track changes to the VM system vs converting to uintptr_t. I believe @bsdjhb thinks we should convert some more of them to void * where we need to preserve provenance, but don't need to be manipulated. Should we upstream vm_pointer_t at all or should we be migrating the VM system away from mach types as part of upstreaming?

[bsdjhb]

I think we should use real pointers when possible, but even after that we will still have some cases left. I'm not yet sure if I prefer using uintptr_t or vm_pointer_t for those cases. I might still prefer vm_pointer_t to mirror vm_offset_t.