[PATCH 3.14 11/11] mm: thp: fix SMP race condition between THP page fault and MADV_DONTNEED

From: Greg Kroah-Hartman
Date: Fri Sep 09 2016 - 11:34:34 EST


3.14-stable review patch. If anyone has any objections, please let me know.

------------------

From: Andrea Arcangeli <aarcange@xxxxxxxxxx>

commit ad33bb04b2a6cee6c1f99fabb15cddbf93ff0433 upstream.

pmd_trans_unstable()/pmd_none_or_trans_huge_or_clear_bad() were
introduced to locklessy (but atomically) detect when a pmd is a regular
(stable) pmd or when the pmd is unstable and can infinitely transition
from pmd_none() and pmd_trans_huge() from under us, while only holding
the mmap_sem for reading (for writing not).

While holding the mmap_sem only for reading, MADV_DONTNEED can run from
under us and so before we can assume the pmd to be a regular stable pmd
we need to compare it against pmd_none() and pmd_trans_huge() in an
atomic way, with pmd_trans_unstable(). The old pmd_trans_huge() left a
tiny window for a race.

Useful applications are unlikely to notice the difference as doing
MADV_DONTNEED concurrently with a page fault would lead to undefined
behavior.

[js] 3.12 backport: no pmd_devmap in 3.12 yet.

[akpm@xxxxxxxxxxxxxxxxxxxx: tidy up comment grammar/layout]
Signed-off-by: Andrea Arcangeli <aarcange@xxxxxxxxxx>
Reported-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
Cc: Vlastimil Babka <vbabka@xxxxxxx>
Signed-off-by: Jiri Slaby <jslaby@xxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>


---
mm/memory.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)

--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3770,8 +3770,18 @@ static int __handle_mm_fault(struct mm_s
if (unlikely(pmd_none(*pmd)) &&
unlikely(__pte_alloc(mm, vma, pmd, address)))
return VM_FAULT_OOM;
- /* if an huge pmd materialized from under us just retry later */
- if (unlikely(pmd_trans_huge(*pmd)))
+ /*
+ * If a huge pmd materialized under us just retry later. Use
+ * pmd_trans_unstable() instead of pmd_trans_huge() to ensure the pmd
+ * didn't become pmd_trans_huge under us and then back to pmd_none, as
+ * a result of MADV_DONTNEED running immediately after a huge pmd fault
+ * in a different thread of this mm, in turn leading to a misleading
+ * pmd_trans_huge() retval. All we have to ensure is that it is a
+ * regular pmd that we can walk with pte_offset_map() and we can do that
+ * through an atomic read in C, which is what pmd_trans_unstable()
+ * provides.
+ */
+ if (unlikely(pmd_trans_unstable(pmd)))
return 0;
/*
* A regular pmd is established and it can't morph into a huge pmd