[4.2.y-ckt stable] Patch "mm: thp: fix SMP race condition between THP page fault and MADV_DONTNEED" has been added to the 4.2.y-ckt tree

[Kamal Mostafa]

This is a note to let you know that I have just added a patch titled

    mm: thp: fix SMP race condition between THP page fault and MADV_DONTNEED

to the linux-4.2.y-queue branch of the 4.2.y-ckt extended stable tree 
which can be found at:

    http://kernel.ubuntu.com/git/ubuntu/linux.git/log/?h=linux-4.2.y-queue

This patch is scheduled to be released in version 4.2.8-ckt5.

If you, or anyone else, feels it should not be added to this tree, please 
reply to this email.

For more information about the 4.2.y-ckt tree, see
https://wiki.ubuntu.com/Kernel/Dev/ExtendedStable

Thanks.
-Kamal

---8<------------------------------------------------------------

>From 7a977a48930352eb23bf87a7e441f5963f311ef8 Mon Sep 17 00:00:00 2001
From: Andrea Arcangeli <aarcange at redhat.com>
Date: Fri, 26 Feb 2016 15:19:28 -0800
Subject: mm: thp: fix SMP race condition between THP page fault and
 MADV_DONTNEED

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.

[akpm at linux-foundation.org: tidy up comment grammar/layout]
Signed-off-by: Andrea Arcangeli <aarcange at redhat.com>
Reported-by: Kirill A. Shutemov <kirill.shutemov at linux.intel.com>
Signed-off-by: Andrew Morton <akpm at linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds at linux-foundation.org>
[ kamal: backport to 4.2-stable: context ]
Signed-off-by: Kamal Mostafa <kamal at canonical.com>
---
 mm/memory.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index 388dcf9..90e6455 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3365,8 +3365,18 @@ static int __handle_mm_fault(struct mm_struct *mm, struct vm_area_struct *vma,
 	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
--
2.7.0