kvm: fix types for `kvm_walk_pages`

[emaste]

Summary

As with other libkvm interfaces use maximum-sized types to support cross-debugging (e.g. a 64-bit vmcore on a 32-bit host). See https://lists.freebsd.org/pipermail/svn-src-all/2019-February/176051.html for further discussion.

[emaste]

sys/sys/types.h

260–261 ↗

(On Diff #63048)

there was some suggestion of introducing kvm_ types -- kvm_paddr_t, kvm_vaddr_t?

[brooks]

sys/sys/types.h

260–261 ↗

(On Diff #63048)

It probably makes sense to have a kpaddr_t in general. I'm not sure how much value lvm_* variants add in practice.

[emaste]

Note that the types are the same size on LP64 archs. This breaks the ABI on ILP32.

[jhb]

sys/sys/types.h

255 ↗

(On Diff #63048)

This comment needs updating, perhaps:

Types suitable for exporting physical addresses, virtual addresses (pointers), and
memory object sizes from the kernel independent of native word size.  These
should be used in place of vm_paddr_t, (u)intptr_t, and size_t in structs which
contain such types that are shared with userspace.

[emaste]

chase member rename

[emaste]

update comment per jhb

[emaste]

update comment again

[will]

I'm fine with these changes, even though they break the API. For my purposes, it isn't a big deal, especially with the version bump.

I don't think I can say much about the appropriateness of the types, but it makes sense to me to have them.