Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull locking changes from Ingo Molnar:
 "The main changes in this cycle were:

   - More gradual enhancements to atomic ops: new atomic*_read_ctrl()
     ops, synchronize atomic_{read,set}() ordering requirements between
     architectures, add atomic_long_t bitops.  (Peter Zijlstra)

   - Add _{relaxed|acquire|release}() variants for inc/dec atomics and
     use them in various locking primitives: mutex, rtmutex, mcs, rwsem.
     This enables weakly ordered architectures (such as arm64) to make
     use of more locking related optimizations.  (Davidlohr Bueso)

   - Implement atomic[64]_{inc,dec}_relaxed() on ARM.  (Will Deacon)

   - Futex kernel data cache footprint micro-optimization.  (Rasmus
     Villemoes)

   - pvqspinlock runtime overhead micro-optimization.  (Waiman Long)

   - misc smaller fixlets"

* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  ARM, locking/atomics: Implement _relaxed variants of atomic[64]_{inc,dec}
  locking/rwsem: Use acquire/release semantics
  locking/mcs: Use acquire/release semantics
  locking/rtmutex: Use acquire/release semantics
  locking/mutex: Use acquire/release semantics
  locking/asm-generic: Add _{relaxed|acquire|release}() variants for inc/dec atomics
  atomic: Implement atomic_read_ctrl()
  atomic, arch: Audit atomic_{read,set}()
  atomic: Add atomic_long_t bitops
  futex: Force hot variables into a single cache line
  locking/pvqspinlock: Kick the PV CPU unconditionally when _Q_SLOW_VAL
  locking/osq: Relax atomic semantics
  locking/qrwlock: Rename ->lock to ->wait_lock
  locking/Documentation/lockstat: Fix typo - lokcing -> locking
  locking/atomics, cmpxchg: Privatize the inclusion of asm/cmpxchg.h

6 files changed, 60 insertions, 43 deletions

diff --git a/include/asm-generic/atomic-long.h b/include/asm-generic/atomic-long.h
index a94cbebbc33d..f91093c2984c 100644
--- a/include/asm-generic/atomic-long.h
+++ b/include/asm-generic/atomic-long.h
@@ -35,7 +35,7 @@ typedef atomic_t atomic_long_t;
 #endif
 
 #define ATOMIC_LONG_READ_OP(mo)						\
-static inline long atomic_long_read##mo(atomic_long_t *l)		\
+static inline long atomic_long_read##mo(const atomic_long_t *l)		\
 {									\
 	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;		\
 									\
@@ -43,6 +43,7 @@ static inline long atomic_long_read##mo(atomic_long_t *l)		\
 }
 ATOMIC_LONG_READ_OP()
 ATOMIC_LONG_READ_OP(_acquire)
+ATOMIC_LONG_READ_OP(_ctrl)
 
 #undef ATOMIC_LONG_READ_OP
 
@@ -112,19 +113,23 @@ static inline void atomic_long_dec(atomic_long_t *l)
 	ATOMIC_LONG_PFX(_dec)(v);
 }
 
-static inline void atomic_long_add(long i, atomic_long_t *l)
-{
-	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;
-
-	ATOMIC_LONG_PFX(_add)(i, v);
+#define ATOMIC_LONG_OP(op)						\
+static inline void							\
+atomic_long_##op(long i, atomic_long_t *l)				\
+{									\
+	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;		\
+									\
+	ATOMIC_LONG_PFX(_##op)(i, v);					\
 }
 
-static inline void atomic_long_sub(long i, atomic_long_t *l)
-{
-	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;
+ATOMIC_LONG_OP(add)
+ATOMIC_LONG_OP(sub)
+ATOMIC_LONG_OP(and)
+ATOMIC_LONG_OP(or)
+ATOMIC_LONG_OP(xor)
+ATOMIC_LONG_OP(andnot)
 
-	ATOMIC_LONG_PFX(_sub)(i, v);
-}
+#undef ATOMIC_LONG_OP
 
 static inline int atomic_long_sub_and_test(long i, atomic_long_t *l)
 {
@@ -154,19 +159,24 @@ static inline int atomic_long_add_negative(long i, atomic_long_t *l)
 	return ATOMIC_LONG_PFX(_add_negative)(i, v);
 }
 
-static inline long atomic_long_inc_return(atomic_long_t *l)
-{
-	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;
-
-	return (long)ATOMIC_LONG_PFX(_inc_return)(v);
-}
-
-static inline long atomic_long_dec_return(atomic_long_t *l)
-{
-	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;
-
-	return (long)ATOMIC_LONG_PFX(_dec_return)(v);
+#define ATOMIC_LONG_INC_DEC_OP(op, mo)					\
+static inline long							\
+atomic_long_##op##_return##mo(atomic_long_t *l)				\
+{									\
+	ATOMIC_LONG_PFX(_t) *v = (ATOMIC_LONG_PFX(_t) *)l;		\
+									\
+	return (long)ATOMIC_LONG_PFX(_##op##_return##mo)(v);		\
 }
+ATOMIC_LONG_INC_DEC_OP(inc,)
+ATOMIC_LONG_INC_DEC_OP(inc, _relaxed)
+ATOMIC_LONG_INC_DEC_OP(inc, _acquire)
+ATOMIC_LONG_INC_DEC_OP(inc, _release)
+ATOMIC_LONG_INC_DEC_OP(dec,)
+ATOMIC_LONG_INC_DEC_OP(dec, _relaxed)
+ATOMIC_LONG_INC_DEC_OP(dec, _acquire)
+ATOMIC_LONG_INC_DEC_OP(dec, _release)
+
+#undef ATOMIC_LONG_INC_DEC_OP
 
 static inline long atomic_long_add_unless(atomic_long_t *l, long a, long u)
 {
diff --git a/include/asm-generic/atomic.h b/include/asm-generic/atomic.h
index d4d7e337fdcb..74f1a3704d7a 100644
--- a/include/asm-generic/atomic.h
+++ b/include/asm-generic/atomic.h
@@ -127,7 +127,7 @@ ATOMIC_OP(xor, ^)
  * Atomically reads the value of @v.
  */
 #ifndef atomic_read
-#define atomic_read(v)	ACCESS_ONCE((v)->counter)
+#define atomic_read(v)	READ_ONCE((v)->counter)
 #endif
 
 /**
@@ -137,7 +137,7 @@ ATOMIC_OP(xor, ^)
  *
  * Atomically sets the value of @v to @i.
  */
-#define atomic_set(v, i) (((v)->counter) = (i))
+#define atomic_set(v, i) WRITE_ONCE(((v)->counter), (i))
 
 #include <linux/irqflags.h>
 
diff --git a/include/asm-generic/mutex-dec.h b/include/asm-generic/mutex-dec.h
index d4f9fb4e53df..fd694cfd678a 100644
--- a/include/asm-generic/mutex-dec.h
+++ b/include/asm-generic/mutex-dec.h
@@ -20,7 +20,7 @@
 static inline void
 __mutex_fastpath_lock(atomic_t *count, void (*fail_fn)(atomic_t *))
 {
-	if (unlikely(atomic_dec_return(count) < 0))
+	if (unlikely(atomic_dec_return_acquire(count) < 0))
 		fail_fn(count);
 }
 
@@ -35,7 +35,7 @@ __mutex_fastpath_lock(atomic_t *count, void (*fail_fn)(atomic_t *))
 static inline int
 __mutex_fastpath_lock_retval(atomic_t *count)
 {
-	if (unlikely(atomic_dec_return(count) < 0))
+	if (unlikely(atomic_dec_return_acquire(count) < 0))
 		return -1;
 	return 0;
 }
@@ -56,7 +56,7 @@ __mutex_fastpath_lock_retval(atomic_t *count)
 static inline void
 __mutex_fastpath_unlock(atomic_t *count, void (*fail_fn)(atomic_t *))
 {
-	if (unlikely(atomic_inc_return(count) <= 0))
+	if (unlikely(atomic_inc_return_release(count) <= 0))
 		fail_fn(count);
 }
 
@@ -80,7 +80,7 @@ __mutex_fastpath_unlock(atomic_t *count, void (*fail_fn)(atomic_t *))
 static inline int
 __mutex_fastpath_trylock(atomic_t *count, int (*fail_fn)(atomic_t *))
 {
-	if (likely(atomic_cmpxchg(count, 1, 0) == 1))
+	if (likely(atomic_cmpxchg_acquire(count, 1, 0) == 1))
 		return 1;
 	return 0;
 }
diff --git a/include/asm-generic/mutex-xchg.h b/include/asm-generic/mutex-xchg.h
index f169ec064785..a6b4a7bd6ac9 100644
--- a/include/asm-generic/mutex-xchg.h
+++ b/include/asm-generic/mutex-xchg.h
@@ -31,7 +31,7 @@ __mutex_fastpath_lock(atomic_t *count, void (*fail_fn)(atomic_t *))
 		 * to ensure that any waiting tasks are woken up by the
 		 * unlock slow path.
 		 */
-		if (likely(atomic_xchg(count, -1) != 1))
+		if (likely(atomic_xchg_acquire(count, -1) != 1))
 			fail_fn(count);
 }
 
@@ -46,7 +46,7 @@ __mutex_fastpath_lock(atomic_t *count, void (*fail_fn)(atomic_t *))
 static inline int
 __mutex_fastpath_lock_retval(atomic_t *count)
 {
-	if (unlikely(atomic_xchg(count, 0) != 1))
+	if (unlikely(atomic_xchg_acquire(count, 0) != 1))
 		if (likely(atomic_xchg(count, -1) != 1))
 			return -1;
 	return 0;
@@ -67,7 +67,7 @@ __mutex_fastpath_lock_retval(atomic_t *count)
 static inline void
 __mutex_fastpath_unlock(atomic_t *count, void (*fail_fn)(atomic_t *))
 {
-	if (unlikely(atomic_xchg(count, 1) != 0))
+	if (unlikely(atomic_xchg_release(count, 1) != 0))
 		fail_fn(count);
 }
 
@@ -91,7 +91,7 @@ __mutex_fastpath_unlock(atomic_t *count, void (*fail_fn)(atomic_t *))
 static inline int
 __mutex_fastpath_trylock(atomic_t *count, int (*fail_fn)(atomic_t *))
 {
-	int prev = atomic_xchg(count, 0);
+	int prev = atomic_xchg_acquire(count, 0);
 
 	if (unlikely(prev < 0)) {
 		/*
@@ -105,7 +105,7 @@ __mutex_fastpath_trylock(atomic_t *count, int (*fail_fn)(atomic_t *))
 		 *   owner's unlock path needlessly, but that's not a problem
 		 *   in practice. ]
 		 */
-		prev = atomic_xchg(count, prev);
+		prev = atomic_xchg_acquire(count, prev);
 		if (prev < 0)
 			prev = 0;
 	}
diff --git a/include/asm-generic/qrwlock_types.h b/include/asm-generic/qrwlock_types.h
index 4d76f24df518..0abc6b6062fb 100644
--- a/include/asm-generic/qrwlock_types.h
+++ b/include/asm-generic/qrwlock_types.h
@@ -10,12 +10,12 @@
 
 typedef struct qrwlock {
 	atomic_t		cnts;
-	arch_spinlock_t		lock;
+	arch_spinlock_t		wait_lock;
 } arch_rwlock_t;
 
 #define	__ARCH_RW_LOCK_UNLOCKED {		\
 	.cnts = ATOMIC_INIT(0),			\
-	.lock = __ARCH_SPIN_LOCK_UNLOCKED,	\
+	.wait_lock = __ARCH_SPIN_LOCK_UNLOCKED,	\
 }
 
 #endif /* __ASM_GENERIC_QRWLOCK_TYPES_H */
diff --git a/include/asm-generic/rwsem.h b/include/asm-generic/rwsem.h
index d48bf5a95cc1..d6d5dc98d7da 100644
--- a/include/asm-generic/rwsem.h
+++ b/include/asm-generic/rwsem.h
@@ -33,7 +33,7 @@
  */
 static inline void __down_read(struct rw_semaphore *sem)
 {
-	if (unlikely(atomic_long_inc_return((atomic_long_t *)&sem->count) <= 0))
+	if (unlikely(atomic_long_inc_return_acquire((atomic_long_t *)&sem->count) <= 0))
 		rwsem_down_read_failed(sem);
 }
 
@@ -42,7 +42,7 @@ static inline int __down_read_trylock(struct rw_semaphore *sem)
 	long tmp;
 
 	while ((tmp = sem->count) >= 0) {
-		if (tmp == cmpxchg(&sem->count, tmp,
+		if (tmp == cmpxchg_acquire(&sem->count, tmp,
 				   tmp + RWSEM_ACTIVE_READ_BIAS)) {
 			return 1;
 		}
@@ -57,7 +57,7 @@ static inline void __down_write_nested(struct rw_semaphore *sem, int subclass)
 {
 	long tmp;
 
-	tmp = atomic_long_add_return(RWSEM_ACTIVE_WRITE_BIAS,
+	tmp = atomic_long_add_return_acquire(RWSEM_ACTIVE_WRITE_BIAS,
 				     (atomic_long_t *)&sem->count);
 	if (unlikely(tmp != RWSEM_ACTIVE_WRITE_BIAS))
 		rwsem_down_write_failed(sem);
@@ -72,7 +72,7 @@ static inline int __down_write_trylock(struct rw_semaphore *sem)
 {
 	long tmp;
 
-	tmp = cmpxchg(&sem->count, RWSEM_UNLOCKED_VALUE,
+	tmp = cmpxchg_acquire(&sem->count, RWSEM_UNLOCKED_VALUE,
 		      RWSEM_ACTIVE_WRITE_BIAS);
 	return tmp == RWSEM_UNLOCKED_VALUE;
 }
@@ -84,7 +84,7 @@ static inline void __up_read(struct rw_semaphore *sem)
 {
 	long tmp;
 
-	tmp = atomic_long_dec_return((atomic_long_t *)&sem->count);
+	tmp = atomic_long_dec_return_release((atomic_long_t *)&sem->count);
 	if (unlikely(tmp < -1 && (tmp & RWSEM_ACTIVE_MASK) == 0))
 		rwsem_wake(sem);
 }
@@ -94,7 +94,7 @@ static inline void __up_read(struct rw_semaphore *sem)
  */
 static inline void __up_write(struct rw_semaphore *sem)
 {
-	if (unlikely(atomic_long_sub_return(RWSEM_ACTIVE_WRITE_BIAS,
+	if (unlikely(atomic_long_sub_return_release(RWSEM_ACTIVE_WRITE_BIAS,
 				 (atomic_long_t *)&sem->count) < 0))
 		rwsem_wake(sem);
 }
@@ -114,7 +114,14 @@ static inline void __downgrade_write(struct rw_semaphore *sem)
 {
 	long tmp;
 
-	tmp = atomic_long_add_return(-RWSEM_WAITING_BIAS,
+	/*
+	 * When downgrading from exclusive to shared ownership,
+	 * anything inside the write-locked region cannot leak
+	 * into the read side. In contrast, anything in the
+	 * read-locked region is ok to be re-ordered into the
+	 * write side. As such, rely on RELEASE semantics.
+	 */
+	tmp = atomic_long_add_return_release(-RWSEM_WAITING_BIAS,
 				     (atomic_long_t *)&sem->count);
 	if (tmp < 0)
 		rwsem_downgrade_wake(sem);