Each page mapped in a process's address space must be correctly
accounted for in _mapcount.  Normally the rules for this are
straightforward but hugetlbfs page table sharing is different.  The page
table pages at the PMD level are reference counted while the mapcount
remains the same.

If this accounting is wrong, it causes bugs like this one reported by
Larry Woodman:

  kernel BUG at mm/filemap.c:135!
  invalid opcode: 0000 [#1] SMP
  CPU 22
  Modules linked in: bridge stp llc sunrpc binfmt_misc dcdbas microcode pcspkr acpi_pad acpi]
  Pid: 18001, comm: mpitest Tainted: G        W    3.3.0+ #4 Dell Inc. PowerEdge R620/07NDJ2
  RIP: 0010:[<ffffffff8112cfed>]  [<ffffffff8112cfed>] __delete_from_page_cache+0x15d/0x170
  Process mpitest (pid: 18001, threadinfo ffff880428972000, task ffff880428b5cc20)
  Call Trace:
    delete_from_page_cache+0x40/0x80
    truncate_hugepages+0x115/0x1f0
    hugetlbfs_evict_inode+0x18/0x30
    evict+0x9f/0x1b0
    iput_final+0xe3/0x1e0
    iput+0x3e/0x50
    d_kill+0xf8/0x110
    dput+0xe2/0x1b0
    __fput+0x162/0x240

During fork(), copy_hugetlb_page_range() detects if huge_pte_alloc()
shared page tables with the check dst_pte == src_pte.  The logic is if
the PMD page is the same, they must be shared.  This assumes that the
sharing is between the parent and child.  However, if the sharing is
with a different process entirely then this check fails as in this
diagram:

  parent
    |
    ------------>pmd
                 src_pte----------> data page
                                        ^
  other--------->pmd--------------------|
                  ^
  child-----------|
                 dst_pte

For this situation to occur, it must be possible for Parent and Other to
have faulted and failed to share page tables with each other.  This is
possible due to the following style of race.

  PROC A                                          PROC B
  copy_hugetlb_page_range                         copy_hugetlb_page_range
    src_pte == huge_pte_offset                      src_pte == huge_pte_offset
    !src_pte so no sharing                          !src_pte so no sharing

  (time passes)

  hugetlb_fault                                   hugetlb_fault
    huge_pte_alloc                                  huge_pte_alloc
      huge_pmd_share                                 huge_pmd_share
        LOCK(i_mmap_mutex)
        find nothing, no sharing
        UNLOCK(i_mmap_mutex)
                                                      LOCK(i_mmap_mutex)
                                                      find nothing, no sharing
                                                      UNLOCK(i_mmap_mutex)
      pmd_alloc                                       pmd_alloc
      LOCK(instantiation_mutex)
      fault
      UNLOCK(instantiation_mutex)
                                                  LOCK(instantiation_mutex)
                                                  fault
                                                  UNLOCK(instantiation_mutex)

These two processes are not poing to the same data page but are not
sharing page tables because the opportunity was missed.  When either
process later forks, the src_pte == dst pte is potentially insufficient.
As the check falls through, the wrong PTE information is copied in
(harmless but wrong) and the mapcount is bumped for a page mapped by a
shared page table leading to the BUG_ON.

This patch addresses the issue by moving pmd_alloc into huge_pmd_share
which guarantees that the shared pud is populated in the same critical
section as pmd.  This also means that huge_pte_offset test in
huge_pmd_share is serialized correctly now which in turn means that the
success of the sharing will be higher as the racing tasks see the pud
and pmd populated together.

Race identified and changelog written mostly by Mel Gorman.

{akpm@linux-foundation.org: attempt to make the huge_pmd_share() comment comprehensible, clean up coding style]
Reported-by: Larry Woodman <lwoodman@redhat.com>
Tested-by: Larry Woodman <lwoodman@redhat.com>
Reviewed-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Michal Hocko <mhocko@suse.cz>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Ken Chen <kenchen@google.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Hillf Danton <dhillf@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull m68knommu arch fixes from Greg Ungerer:
 "This contains 2 fixes.  One fixes compilation of ColdFire clk code,
  the other makes sure we use the generic atomic64 support on all m68k
  targets."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu:
  m68k: select CONFIG_GENERIC_ATOMIC64 for all m68k CPU types
  m68knommu: select CONFIG_HAVE_CLK for ColdFire CPU types

Pull config cleanup for ia64 from Tony Luck:
 "Clean out references to dead CONFIG_MISC_DEVICES option"

* tag 'please-pull-ia64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux:
  [IA64] defconfig: Remove CONFIG_MISC_DEVICES

commit 7c5763b845 (drivers:misc: Remove MISC_DEVICES config option) removed
CONFIG_MISC_DEVICES option, so remove the occurrences from the config files
as well.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>

Pull x86 fixes from Ingo Molnar.

A x32 socket ABI fix with a -stable backport tag among other fixes.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x32: Use compat shims for {g,s}etsockopt
  Revert "x86-64/efi: Use EFI to deal with platform wall clock"
  x86, apic: fix broken legacy interrupts in the logical apic mode
  x86, build: Globally set -fno-pic
  x86, avx: don't use avx instructions with "noxsave" boot param

Pull x86 perf fixes from Ingo Molnar.

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86: disable PEBS on a guest entry.
  perf/x86: Add Intel Westmere-EX uncore support
  perf/x86: Fixes for Nehalem-EX uncore driver
  perf, x86: Fix uncore_types_exit section mismatch

Merge alpha architecture update from Michael Cree:
 "The Alpha Maintainer, Matt Turner, is currently unavailable, so I have
  collected up patches that have been posted to the linux-alpha mailing
  list over the last couple of months, and are forwarding them to you in
  the hope that you are prepared to accept them via me.

  The patches by Al Viro and myself I have been running against kernels
  for two months now so have had quite a bit of testing.  All except one
  patch were intended for the 3.5 kernel but because of Matt's
  unavailability never got forwarded to you."

* emailed patches from Michael Cree <mcree@orcon.net.nz>: (9 commits)
  alpha: Fix fall-out from disintegrating asm/system.h
  Redefine ATOMIC_INIT and ATOMIC64_INIT to drop the casts
  alpha: fix fpu.h usage in userspace
  alpha/mm/fault.c: Port OOM changes to do_page_fault
  alpha: take kernel_execve() out of entry.S
  alpha: take a bunch of syscalls into osf_sys.c
  alpha: Use new generic strncpy_from_user() and strnlen_user()
  alpha: Wire up cross memory attach syscalls
  alpha: Don't export SOCK_NONBLOCK to user space.

Commit ec2212088c42 ("Disintegrate asm/system.h for Alpha") removed
asm/system.h however arch/alpha/oprofile/common.c requires definitions
that were shifted from asm/system.h to asm/special_insns.h.  Include
that.

Signed-off-by: Michael Cree <mcree@orcon.net.nz>
Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The following build error occurred during an alpha build:

  net/core/sock.c:274:36: error: initializer element is not constant

Dave Anglin says:
> Here is the line in sock.i:
>
> struct static_key memalloc_socks = ((struct static_key) { .enabled =
> ((atomic_t) { (0) }) });

The above line contains two compound literals.  It also uses a designated
initializer to initialize the field enabled.  A compound literal is not a
constant expression.

The location of the above statement isn't fully clear, but if a compound
literal occurs outside the body of a function, the initializer list must
consist of constant expressions.

Cc: <stable@vger.kernel.org>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Michael Cree <mcree@orcon.net.nz>
Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

After commit ec2212088c42 ("Disintegrate asm/system.h for Alpha"), the
fpu.h header which we install for userland started depending on
special_insns.h which is not installed.

However, fpu.h only uses that for __KERNEL__ code, so protect the
inclusion the same way to avoid build breakage in glibc:

  /usr/include/asm/fpu.h:4:31: fatal error: asm/special_insns.h: No such file or directory

Cc: stable@vger.kernel.org
Reported-by: Matt Turner <mattst88@gentoo.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Michael Cree <mcree@orcon.net.nz>
Acked-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>