[PATCH 4.4 31/87] futex: Replace barrier() in unqueue_me() with READ_ONCE()

From: Greg Kroah-Hartman
Date: Mon Jan 15 2018 - 07:39:39 EST


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

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

From: Jianyu Zhan <nasa4836@xxxxxxxxx>

commit 29b75eb2d56a714190a93d7be4525e617591077a upstream.

Commit e91467ecd1ef ("bug in futex unqueue_me") introduced a barrier() in
unqueue_me() to prevent the compiler from rereading the lock pointer which
might change after a check for NULL.

Replace the barrier() with a READ_ONCE() for the following reasons:

1) READ_ONCE() is a weaker form of barrier() that affects only the specific
load operation, while barrier() is a general compiler level memory barrier.
READ_ONCE() was not available at the time when the barrier was added.

2) Aside of that READ_ONCE() is descriptive and self explainatory while a
barrier without comment is not clear to the casual reader.

No functional change.

[ tglx: Massaged changelog ]

Signed-off-by: Jianyu Zhan <nasa4836@xxxxxxxxx>
Acked-by: Christian Borntraeger <borntraeger@xxxxxxxxxx>
Acked-by: Darren Hart <dvhart@xxxxxxxxxxxxxxx>
Cc: dave@xxxxxxxxxxxx
Cc: peterz@xxxxxxxxxxxxx
Cc: linux@xxxxxxxxxxxxxxxxxx
Cc: akpm@xxxxxxxxxxxxxxxxxxxx
Cc: fengguang.wu@xxxxxxxxx
Cc: bigeasy@xxxxxxxxxxxxx
Link: http://lkml.kernel.org/r/1457314344-5685-1-git-send-email-nasa4836@xxxxxxxxx
Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Signed-off-by: Davidlohr Bueso <dbueso@xxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
kernel/futex.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)

--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -1939,8 +1939,12 @@ static int unqueue_me(struct futex_q *q)

/* In the common case we don't take the spinlock, which is nice. */
retry:
- lock_ptr = q->lock_ptr;
- barrier();
+ /*
+ * q->lock_ptr can change between this read and the following spin_lock.
+ * Use READ_ONCE to forbid the compiler from reloading q->lock_ptr and
+ * optimizing lock_ptr out of the logic below.
+ */
+ lock_ptr = READ_ONCE(q->lock_ptr);
if (lock_ptr != NULL) {
spin_lock(lock_ptr);
/*