[PATCH V5 0/3] THP related cleanups

["Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>]

Changes from V4:
* Folded patches in -mm
  mm-thp-split-out-pmd-collpase-flush-into-a-separate-functions-fix.patch
  mm-thp-split-out-pmd-collpase-flush-into-a-separate-functions-fix-2.patch
  mm-clarify-that-the-function-operateds-on-hugepage-pte-fix.patch
* Fix VM_BUG_ON on x86.
 the default implementation of pmdp_collapse_flush used the hugepage variant
 and hence can be called on pmd_t pointing to pgtable. This resulting in us
 hitting VM_BUG_ON in pmdp_clear_flush. Update powerpc/mm: Use generic version of pmdp_clear_flush
 to handle this.


NOTE: Can we get this tested on s390 ?

Aneesh Kumar K.V (3):
  mm/thp: Split out pmd collpase flush into a separate functions
  powerpc/mm: Use generic version of pmdp_clear_flush
  mm: Clarify that the function operates on hugepage pte

 arch/mips/include/asm/pgtable.h          |  8 ++--
 arch/powerpc/include/asm/pgtable-ppc64.h | 14 +++---
 arch/powerpc/mm/pgtable_64.c             | 73 +++++++++++++++-----------------
 arch/s390/include/asm/pgtable.h          | 30 ++++++++-----
 arch/sparc/include/asm/pgtable_64.h      |  8 ++--
 arch/tile/include/asm/pgtable.h          |  8 ++--
 arch/x86/include/asm/pgtable.h           |  4 +-
 include/asm-generic/pgtable.h            | 49 +++++++++++++++++----
 include/linux/mmu_notifier.h             | 12 +++---
 mm/huge_memory.c                         | 18 ++++----
 mm/migrate.c                             |  2 +-
 mm/pgtable-generic.c                     |  9 ++--
 mm/rmap.c                                |  2 +-
 13 files changed, 136 insertions(+), 101 deletions(-)

-- 
2.1.4