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

Pull core locking updates from Ingo Molnar:
 "The main changes are:

   - mutex, completions and rtmutex micro-optimizations
   - lock debugging fix
   - various cleanups in the MCS and the futex code"

* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  locking/rtmutex: Optimize setting task running after being blocked
  locking/rwsem: Use task->state helpers
  sched/completion: Add lock-free checking of the blocking case
  sched/completion: Remove unnecessary ->wait.lock serialization when reading completion state
  locking/mutex: Explicitly mark task as running after wakeup
  futex: Fix argument handling in futex_lock_pi() calls
  doc: Fix misnamed FUTEX_CMP_REQUEUE_PI op constants
  locking/Documentation: Update code path
  softirq/preempt: Add missing current->preempt_disable_ip update
  locking/osq: No need for load/acquire when acquire-polling
  locking/mcs: Better differentiate between MCS variants
  locking/mutex: Introduce ww_mutex_set_context_slowpath()
  locking/mutex: Move MCS related comments to proper location
  locking/mutex: Checking the stamp is WW only

Documentation/futex-requeue-pi.txt

@@ -98,7 +98,7 @@ rt_mutex_start_proxy_lock() and rt_mutex_finish_proxy_lock(), which
 allow the requeue code to acquire an uncontended rt_mutex on behalf
 of the waiter and to enqueue the waiter on a contended rt_mutex.
 Two new system calls provide the kernel<->user interface to
-requeue_pi: FUTEX_WAIT_REQUEUE_PI and FUTEX_REQUEUE_CMP_PI.
+requeue_pi: FUTEX_WAIT_REQUEUE_PI and FUTEX_CMP_REQUEUE_PI.
 
 FUTEX_WAIT_REQUEUE_PI is called by the waiter (pthread_cond_wait()
 and pthread_cond_timedwait()) to block on the initial futex and wait
@@ -107,7 +107,7 @@ result of a high-speed collision between futex_wait() and
 futex_lock_pi(), with some extra logic to check for the additional
 wake-up scenarios.
 
-FUTEX_REQUEUE_CMP_PI is called by the waker
+FUTEX_CMP_REQUEUE_PI is called by the waker
 (pthread_cond_broadcast() and pthread_cond_signal()) to requeue and
 possibly wake the waiting tasks. Internally, this system call is
 still handled by futex_requeue (by passing requeue_pi=1).  Before
@@ -120,12 +120,12 @@ task as a waiter on the underlying rt_mutex.  It is possible that
 the lock can be acquired at this stage as well, if so, the next
 waiter is woken to finish the acquisition of the lock.
 
-FUTEX_REQUEUE_PI accepts nr_wake and nr_requeue as arguments, but
+FUTEX_CMP_REQUEUE_PI accepts nr_wake and nr_requeue as arguments, but
 their sum is all that really matters.  futex_requeue() will wake or
 requeue up to nr_wake + nr_requeue tasks.  It will wake only as many
 tasks as it can acquire the lock for, which in the majority of cases
 should be 0 as good programming practice dictates that the caller of
 either pthread_cond_broadcast() or pthread_cond_signal() acquire the
-mutex prior to making the call. FUTEX_REQUEUE_PI requires that
+mutex prior to making the call. FUTEX_CMP_REQUEUE_PI requires that
 nr_wake=1.  nr_requeue should be INT_MAX for broadcast and 0 for
 signal.

Documentation/locking/lockdep-design.txt

@@ -34,7 +34,7 @@ The validator tracks lock-class usage history into 4n + 1 separate state bits:
 - 'ever held with STATE enabled'
 - 'ever held as readlock with STATE enabled'
 
-Where STATE can be either one of (kernel/lockdep_states.h)
+Where STATE can be either one of (kernel/locking/lockdep_states.h)
  - hardirq
  - softirq
  - reclaim_fs

include/linux/osq_lock.h

@@ -5,8 +5,11 @@
  * An MCS like lock especially tailored for optimistic spinning for sleeping
  * lock implementations (mutex, rwsem, etc).
  */
-
-#define OSQ_UNLOCKED_VAL (0)
+struct optimistic_spin_node {
+	struct optimistic_spin_node *next, *prev;
+	int locked; /* 1 if lock acquired */
+	int cpu; /* encoded CPU # + 1 value */
+};
 
 struct optimistic_spin_queue {
 	/*
@@ -16,6 +19,8 @@ struct optimistic_spin_queue {
 	atomic_t tail;
 };
 
+#define OSQ_UNLOCKED_VAL (0)
+
 /* Init macro and function. */
 #define OSQ_LOCK_UNLOCKED { ATOMIC_INIT(OSQ_UNLOCKED_VAL) }
 
@@ -24,4 +29,7 @@ static inline void osq_lock_init(struct optimistic_spin_queue *lock)
 	atomic_set(&lock->tail, OSQ_UNLOCKED_VAL);
 }
 
+extern bool osq_lock(struct optimistic_spin_queue *lock);
+extern void osq_unlock(struct optimistic_spin_queue *lock);
+
 #endif

kernel/Kconfig.locks

@@ -231,6 +231,10 @@ config RWSEM_SPIN_ON_OWNER
        def_bool y
        depends on SMP && RWSEM_XCHGADD_ALGORITHM && ARCH_SUPPORTS_ATOMIC_RMW
 
+config LOCK_SPIN_ON_OWNER
+       def_bool y
+       depends on MUTEX_SPIN_ON_OWNER || RWSEM_SPIN_ON_OWNER
+
 config ARCH_USE_QUEUE_RWLOCK
 	bool
 

kernel/futex.c

@@ -2258,7 +2258,7 @@ static long futex_wait_restart(struct restart_block *restart)
  * if there are waiters then it will block, it does PI, etc. (Due to
  * races the kernel might see a 0 value of the futex too.)
  */
-static int futex_lock_pi(u32 __user *uaddr, unsigned int flags, int detect,
+static int futex_lock_pi(u32 __user *uaddr, unsigned int flags,
 			 ktime_t *time, int trylock)
 {
 	struct hrtimer_sleeper timeout, *to = NULL;
@@ -2953,11 +2953,11 @@ long do_futex(u32 __user *uaddr, int op, u32 val, ktime_t *timeout,
 	case FUTEX_WAKE_OP:
 		return futex_wake_op(uaddr, flags, uaddr2, val, val2, val3);
 	case FUTEX_LOCK_PI:
-		return futex_lock_pi(uaddr, flags, val, timeout, 0);
+		return futex_lock_pi(uaddr, flags, timeout, 0);
 	case FUTEX_UNLOCK_PI:
 		return futex_unlock_pi(uaddr, flags);
 	case FUTEX_TRYLOCK_PI:
-		return futex_lock_pi(uaddr, flags, 0, timeout, 1);
+		return futex_lock_pi(uaddr, flags, NULL, 1);
 	case FUTEX_WAIT_REQUEUE_PI:
 		val3 = FUTEX_BITSET_MATCH_ANY;
 		return futex_wait_requeue_pi(uaddr, flags, val, timeout, val3,

kernel/locking/Makefile

@@ -1,5 +1,5 @@
 
-obj-y += mutex.o semaphore.o rwsem.o mcs_spinlock.o
+obj-y += mutex.o semaphore.o rwsem.o
 
 ifdef CONFIG_FUNCTION_TRACER
 CFLAGS_REMOVE_lockdep.o = -pg
@@ -14,6 +14,7 @@ ifeq ($(CONFIG_PROC_FS),y)
 obj-$(CONFIG_LOCKDEP) += lockdep_proc.o
 endif
 obj-$(CONFIG_SMP) += spinlock.o
+obj-$(CONFIG_LOCK_SPIN_ON_OWNER) += osq_lock.o
 obj-$(CONFIG_SMP) += lglock.o
 obj-$(CONFIG_PROVE_LOCKING) += spinlock.o
 obj-$(CONFIG_RT_MUTEXES) += rtmutex.o

kernel/locking/mcs_spinlock.h

@@ -108,20 +108,4 @@ void mcs_spin_unlock(struct mcs_spinlock **lock, struct mcs_spinlock *node)
 	arch_mcs_spin_unlock_contended(&next->locked);
 }
 
-/*
- * Cancellable version of the MCS lock above.
- *
- * Intended for adaptive spinning of sleeping locks:
- * mutex_lock()/rwsem_down_{read,write}() etc.
- */
-
-struct optimistic_spin_node {
-	struct optimistic_spin_node *next, *prev;
-	int locked; /* 1 if lock acquired */
-	int cpu; /* encoded CPU # value */
-};
-
-extern bool osq_lock(struct optimistic_spin_queue *lock);
-extern void osq_unlock(struct optimistic_spin_queue *lock);
-
 #endif /* __LINUX_MCS_SPINLOCK_H */

kernel/locking/mutex.c

@@ -147,7 +147,7 @@ static __always_inline void ww_mutex_lock_acquired(struct ww_mutex *ww,
 }
 
 /*
- * after acquiring lock with fastpath or when we lost out in contested
+ * After acquiring lock with fastpath or when we lost out in contested
  * slowpath, set ctx and wake up any waiters so they can recheck.
  *
  * This function is never called when CONFIG_DEBUG_LOCK_ALLOC is set,
@@ -191,19 +191,32 @@ ww_mutex_set_context_fastpath(struct ww_mutex *lock,
 	spin_unlock_mutex(&lock->base.wait_lock, flags);
 }
 
+/*
+ * After acquiring lock in the slowpath set ctx and wake up any
+ * waiters so they can recheck.
+ *
+ * Callers must hold the mutex wait_lock.
+ */
+static __always_inline void
+ww_mutex_set_context_slowpath(struct ww_mutex *lock,
+			      struct ww_acquire_ctx *ctx)
+{
+	struct mutex_waiter *cur;
+
+	ww_mutex_lock_acquired(lock, ctx);
+	lock->ctx = ctx;
+
+	/*
+	 * Give any possible sleeping processes the chance to wake up,
+	 * so they can recheck if they have to back off.
+	 */
+	list_for_each_entry(cur, &lock->base.wait_list, list) {
+		debug_mutex_wake_waiter(&lock->base, cur);
+		wake_up_process(cur->task);
+	}
+}
 
 #ifdef CONFIG_MUTEX_SPIN_ON_OWNER
-/*
- * In order to avoid a stampede of mutex spinners from acquiring the mutex
- * more or less simultaneously, the spinners need to acquire a MCS lock
- * first before spinning on the owner field.
- *
- */
-
-/*
- * Mutex spinning code migrated from kernel/sched/core.c
- */
-
 static inline bool owner_running(struct mutex *lock, struct task_struct *owner)
 {
 	if (lock->owner != owner)
@@ -307,6 +320,11 @@ static bool mutex_optimistic_spin(struct mutex *lock,
 	if (!mutex_can_spin_on_owner(lock))
 		goto done;
 
+	/*
+	 * In order to avoid a stampede of mutex spinners trying to
+	 * acquire the mutex all at once, the spinners need to take a
+	 * MCS (queued) lock first before spinning on the owner field.
+	 */
 	if (!osq_lock(&lock->osq))
 		goto done;
 
@@ -469,7 +487,7 @@ void __sched ww_mutex_unlock(struct ww_mutex *lock)
 EXPORT_SYMBOL(ww_mutex_unlock);
 
 static inline int __sched
-__mutex_lock_check_stamp(struct mutex *lock, struct ww_acquire_ctx *ctx)
+__ww_mutex_lock_check_stamp(struct mutex *lock, struct ww_acquire_ctx *ctx)
 {
 	struct ww_mutex *ww = container_of(lock, struct ww_mutex, base);
 	struct ww_acquire_ctx *hold_ctx = ACCESS_ONCE(ww->ctx);
@@ -557,7 +575,7 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass,
 		}
 
 		if (use_ww_ctx && ww_ctx->acquired > 0) {
-			ret = __mutex_lock_check_stamp(lock, ww_ctx);
+			ret = __ww_mutex_lock_check_stamp(lock, ww_ctx);
 			if (ret)
 				goto err;
 		}
@@ -569,6 +587,8 @@ __mutex_lock_common(struct mutex *lock, long state, unsigned int subclass,
 		schedule_preempt_disabled();
 		spin_lock_mutex(&lock->wait_lock, flags);
 	}
+	__set_task_state(task, TASK_RUNNING);
+
 	mutex_remove_waiter(lock, &waiter, current_thread_info());
 	/* set it to 0 if there are no waiters left: */
 	if (likely(list_empty(&lock->wait_list)))
@@ -582,23 +602,7 @@ skip_wait:
 
 	if (use_ww_ctx) {
 		struct ww_mutex *ww = container_of(lock, struct ww_mutex, base);
-		struct mutex_waiter *cur;
-
-		/*
-		 * This branch gets optimized out for the common case,
-		 * and is only important for ww_mutex_lock.
-		 */
-		ww_mutex_lock_acquired(ww, ww_ctx);
-		ww->ctx = ww_ctx;
-
-		/*
-		 * Give any possible sleeping processes the chance to wake up,
-		 * so they can recheck if they have to back off.
-		 */
-		list_for_each_entry(cur, &lock->wait_list, list) {
-			debug_mutex_wake_waiter(lock, cur);
-			wake_up_process(cur->task);
-		}
+		ww_mutex_set_context_slowpath(ww, ww_ctx);
 	}
 
 	spin_unlock_mutex(&lock->wait_lock, flags);

kernel/locking/osq_lock.c

@@ -1,8 +1,6 @@
 #include <linux/percpu.h>
 #include <linux/sched.h>
-#include "mcs_spinlock.h"
-
-#ifdef CONFIG_SMP
+#include <linux/osq_lock.h>
 
 /*
  * An MCS like lock especially tailored for optimistic spinning for sleeping
@@ -111,7 +109,7 @@ bool osq_lock(struct optimistic_spin_queue *lock)
 	 * cmpxchg in an attempt to undo our queueing.
 	 */
 
-	while (!smp_load_acquire(&node->locked)) {
+	while (!ACCESS_ONCE(node->locked)) {
 		/*
 		 * If we need to reschedule bail... so we can block.
 		 */
@@ -203,6 +201,3 @@ void osq_unlock(struct optimistic_spin_queue *lock)
 	if (next)
 		ACCESS_ONCE(next->locked) = 1;
 }
-
-#endif
-

kernel/locking/rtmutex.c

@@ -1130,6 +1130,7 @@ __rt_mutex_slowlock(struct rt_mutex *lock, int state,
 		set_current_state(state);
 	}
 
+	__set_current_state(TASK_RUNNING);
 	return ret;
 }
 
@@ -1188,10 +1189,9 @@ rt_mutex_slowlock(struct rt_mutex *lock, int state,
 	ret = task_blocks_on_rt_mutex(lock, &waiter, current, chwalk);
 
 	if (likely(!ret))
+		/* sleep on the mutex */
 		ret = __rt_mutex_slowlock(lock, state, timeout, &waiter);
 
-	set_current_state(TASK_RUNNING);
-
 	if (unlikely(ret)) {
 		remove_waiter(lock, &waiter);
 		rt_mutex_handle_deadlock(ret, chwalk, &waiter);
@@ -1626,10 +1626,9 @@ int rt_mutex_finish_proxy_lock(struct rt_mutex *lock,
 
 	set_current_state(TASK_INTERRUPTIBLE);
 
+	/* sleep on the mutex */
 	ret = __rt_mutex_slowlock(lock, TASK_INTERRUPTIBLE, to, waiter);
 
-	set_current_state(TASK_RUNNING);
-
 	if (unlikely(ret))
 		remove_waiter(lock, waiter);
 

kernel/locking/rwsem-spinlock.c

@@ -154,7 +154,7 @@ void __sched __down_read(struct rw_semaphore *sem)
 		set_task_state(tsk, TASK_UNINTERRUPTIBLE);
 	}
 
-	tsk->state = TASK_RUNNING;
+	__set_task_state(tsk, TASK_RUNNING);
  out:
 	;
 }

kernel/locking/rwsem-xadd.c

@@ -242,8 +242,7 @@ struct rw_semaphore __sched *rwsem_down_read_failed(struct rw_semaphore *sem)
 		schedule();
 	}
 
-	tsk->state = TASK_RUNNING;
-
+	__set_task_state(tsk, TASK_RUNNING);
 	return sem;
 }
 EXPORT_SYMBOL(rwsem_down_read_failed);

kernel/sched/completion.c

@@ -268,6 +268,15 @@ bool try_wait_for_completion(struct completion *x)
 	unsigned long flags;
 	int ret = 1;
 
+	/*
+	 * Since x->done will need to be locked only
+	 * in the non-blocking case, we check x->done
+	 * first without taking the lock so we can
+	 * return early in the blocking case.
+	 */
+	if (!ACCESS_ONCE(x->done))
+		return 0;
+
 	spin_lock_irqsave(&x->wait.lock, flags);
 	if (!x->done)
 		ret = 0;
@@ -288,13 +297,6 @@ EXPORT_SYMBOL(try_wait_for_completion);
  */
 bool completion_done(struct completion *x)
 {
-	unsigned long flags;
-	int ret = 1;
-
-	spin_lock_irqsave(&x->wait.lock, flags);
-	if (!x->done)
-		ret = 0;
-	spin_unlock_irqrestore(&x->wait.lock, flags);
-	return ret;
+	return !!ACCESS_ONCE(x->done);
 }
 EXPORT_SYMBOL(completion_done);

kernel/softirq.c

@@ -114,8 +114,12 @@ void __local_bh_disable_ip(unsigned long ip, unsigned int cnt)
 		trace_softirqs_off(ip);
 	raw_local_irq_restore(flags);
 
-	if (preempt_count() == cnt)
+	if (preempt_count() == cnt) {
+#ifdef CONFIG_DEBUG_PREEMPT
+		current->preempt_disable_ip = get_parent_ip(CALLER_ADDR1);
+#endif
 		trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1));
+	}
 }
 EXPORT_SYMBOL(__local_bh_disable_ip);
 #endif /* CONFIG_TRACE_IRQFLAGS */