Function vm_page_acquire_unlocked both looks up pages and validates them. Much of it serves the needs of only one caller, vm_page_grab_pages_unlocked. To unburden the other callers from that, extract the validation part into a function vm_page_validate(), and let the callers do their own lookups. The change saves all callers a bit of null-checking, and appears to make the amd64 code a tiny bit smaller. It allows future changes to how multiple pages are loaded locklessly to not interfere with how single pages are loaded locklessly.