mm: remove remaining references to NUMA hinting bits and helpers
authorMel Gorman <mgorman@suse.de>
Thu, 12 Feb 2015 22:58:32 +0000 (14:58 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 13 Feb 2015 02:54:08 +0000 (18:54 -0800)
commit21d9ee3eda7792c45880b2f11bff8e95c9a061fb
tree2e20932b8f1526e6d1f48add9e818ed43d7be8ee
parent4d9424669946532be754a6e116618dcb58430cb4
mm: remove remaining references to NUMA hinting bits and helpers

This patch removes the NUMA PTE bits and associated helpers.  As a
side-effect it increases the maximum possible swap space on x86-64.

One potential source of problems is races between the marking of PTEs
PROT_NONE, NUMA hinting faults and migration.  It must be guaranteed that
a PTE being protected is not faulted in parallel, seen as a pte_none and
corrupting memory.  The base case is safe but transhuge has problems in
the past due to an different migration mechanism and a dependance on page
lock to serialise migrations and warrants a closer look.

task_work hinting update parallel fault
------------------------ --------------
change_pmd_range
  change_huge_pmd
    __pmd_trans_huge_lock
      pmdp_get_and_clear
__handle_mm_fault
pmd_none
  do_huge_pmd_anonymous_page
  read? pmd_lock blocks until hinting complete, fail !pmd_none test
  write? __do_huge_pmd_anonymous_page acquires pmd_lock, checks pmd_none
      pmd_modify
      set_pmd_at

task_work hinting update parallel migration
------------------------ ------------------
change_pmd_range
  change_huge_pmd
    __pmd_trans_huge_lock
      pmdp_get_and_clear
__handle_mm_fault
  do_huge_pmd_numa_page
    migrate_misplaced_transhuge_page
    pmd_lock waits for updates to complete, recheck pmd_same
      pmd_modify
      set_pmd_at

Both of those are safe and the case where a transhuge page is inserted
during a protection update is unchanged.  The case where two processes try
migrating at the same time is unchanged by this series so should still be
ok.  I could not find a case where we are accidentally depending on the
PTE not being cleared and flushed.  If one is missed, it'll manifest as
corruption problems that start triggering shortly after this series is
merged and only happen when NUMA balancing is enabled.

Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Sasha Levin <sasha.levin@oracle.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Dave Jones <davej@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Kirill Shutemov <kirill.shutemov@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
arch/powerpc/include/asm/pgtable.h
arch/powerpc/include/asm/pte-common.h
arch/powerpc/include/asm/pte-hash64.h
arch/x86/include/asm/pgtable.h
arch/x86/include/asm/pgtable_64.h
arch/x86/include/asm/pgtable_types.h
include/asm-generic/pgtable.h
include/linux/swapops.h