vm_page_replace: add wrapper to KASSERT about old page

[rlibby]

Summary

It turns out the callers of vm_page_replace know exactly which page they
are replacing and would like to assert about it. Change those from hard
panics to KASSERTs, and provide them with a wrapper so they don't have
to deal with warnings from an INVARIANTS-dependent dead store of the
return value of vm_page_replace.

Preliminary work for D4478.

I'm definitely open to suggestions for renaming the wrapper and
vm_page_replace if you have any ideas.

Test Plan

kyua test -k /usr/tests/sys/Kyuafile

[kib]

Traditionally we used names like XXX_locked(), might be name it vm_page_replace_checked(). IMO the _check/checked suffix is telling enough.

[rlibby]

vm_page_replace: kib feedback: s/check/checked/

[cem]

sys/vm/sg_pager.c

190–193

sg_pager_getpages has been rototilled out from underneath this patch, needs rebasing before I can apply it.

[rlibby]

Rebase, resolving conflict.

Resolution was trivial, I don't think this needs re-review, but I'm
unsure if there's some kind of policy for that.

I am having environment issues so I build-tested the fix up but didn't
run any tests.