]> git.baikalelectronics.ru Git - kernel.git/commit
mm/huge_memory: splitting set mapping+index before unfreeze
authorHugh Dickins <hughd@google.com>
Fri, 30 Nov 2018 22:10:16 +0000 (14:10 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 30 Nov 2018 22:56:15 +0000 (14:56 -0800)
commitd4befe81bd7d99e5409689dadcbf0e27b516e27a
tree88c2c654d6b0c60e437e814e217d520ebd1b5be6
parente81ab4941259e28080b30ea1a2d57f71d45c5e64
mm/huge_memory: splitting set mapping+index before unfreeze

Huge tmpfs stress testing has occasionally hit shmem_undo_range()'s
VM_BUG_ON_PAGE(page_to_pgoff(page) != index, page).

Move the setting of mapping and index up before the page_ref_unfreeze()
in __split_huge_page_tail() to fix this: so that a page cache lookup
cannot get a reference while the tail's mapping and index are unstable.

In fact, might as well move them up before the smp_wmb(): I don't see an
actual need for that, but if I'm missing something, this way round is
safer than the other, and no less efficient.

You might argue that VM_BUG_ON_PAGE(page_to_pgoff(page) != index, page) is
misplaced, and should be left until after the trylock_page(); but left as
is has not crashed since, and gives more stringent assurance.

Link: http://lkml.kernel.org/r/alpine.LSU.2.11.1811261516380.2275@eggly.anvils
Fixes: 171b3a5e1817a ("thp: reintroduce split_huge_page()")
Requires: 1d399aa596fe ("mm/huge_memory.c: reorder operations in __split_huge_page_tail()")
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: <stable@vger.kernel.org> [4.8+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/huge_memory.c