Else after first failing tests all furhter test can't start because
trying to stat testmntpt/.wcfs gives ENOTCONN.

This allows to make decision to drop a page right after writeout for
wcfs case. If vmas for one fileh were allowed to be mixed - we could not
know whether to drop the page & its memory from RAM or not - even if
there is no vma mapped in !wcfs mode, it could be created later, and by
dropping RAM cahce we hit severe slowness. NOTE: in between transactions
there is usually no vmas mapped, so this case is not artificial.

By requiring that all VMAs are of the same kind under one fileh we avoid
this ambiguity.

* master:
  include: Fix typos

* master:
  bigfile: Fix typos

* master:
  bigfile/virtmem/test: Factor-out asserting on ram->lru_list and fileh->dirty_pages
  bigfile/virtmem/test: Fix test_file_access_pagefault's undef place

-> into CHECK_MRU and CHECK_DIRTY.

Besides improving signal/noise ratio in tests, it also gives more
details by printing full lists when things are found to be different
from expected.

The undef defined in that function were placed in next function going
after it which does not define any macro.

* master:
  bigfile/tests: Fix typos