locking: Apply contention tracepoints in the slow path

Adding the lock contention tracepoints in various lock function slow
paths.  Note that each arch can define spinlock differently, I only
added it only to the generic qspinlock for now.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Link: https://lkml.kernel.org/r/20220322185709.141236-3-namhyung@kernel.org

kernel/locking/mutex.c

@@ -644,6 +644,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas
 	}
 
 	set_current_state(state);
+	trace_contention_begin(lock, 0);
 	for (;;) {
 		bool first;
 
@@ -710,6 +711,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas
 skip_wait:
 	/* got the lock - cleanup and rejoice! */
 	lock_acquired(&lock->dep_map, ip);
+	trace_contention_end(lock, 0);
 
 	if (ww_ctx)
 		ww_mutex_lock_acquired(ww, ww_ctx);
@@ -721,6 +723,7 @@ __mutex_lock_common(struct mutex *lock, unsigned int state, unsigned int subclas
 err:
 	__set_current_state(TASK_RUNNING);
 	__mutex_remove_waiter(lock, &waiter);
+	trace_contention_end(lock, ret);
 err_early_kill:
 	raw_spin_unlock(&lock->wait_lock);
 	debug_mutex_free_waiter(&waiter);

kernel/locking/percpu-rwsem.c

@@ -9,6 +9,7 @@
 #include <linux/sched/task.h>
 #include <linux/sched/debug.h>
 #include <linux/errno.h>
+#include <trace/events/lock.h>
 
 int __percpu_init_rwsem(struct percpu_rw_semaphore *sem,
 			const char *name, struct lock_class_key *key)
@@ -171,9 +172,11 @@ bool __sched __percpu_down_read(struct percpu_rw_semaphore *sem, bool try)
 	if (try)
 		return false;
 
+	trace_contention_begin(sem, LCB_F_PERCPU | LCB_F_READ);
 	preempt_enable();
 	percpu_rwsem_wait(sem, /* .reader = */ true);
 	preempt_disable();
+	trace_contention_end(sem, 0);
 
 	return true;
 }
@@ -216,6 +219,7 @@ void __sched percpu_down_write(struct percpu_rw_semaphore *sem)
 {
 	might_sleep();
 	rwsem_acquire(&sem->dep_map, 0, 0, _RET_IP_);
+	trace_contention_begin(sem, LCB_F_PERCPU | LCB_F_WRITE);
 
 	/* Notify readers to take the slow path. */
 	rcu_sync_enter(&sem->rss);
@@ -237,6 +241,7 @@ void __sched percpu_down_write(struct percpu_rw_semaphore *sem)
 
 	/* Wait for all active readers to complete. */
 	rcuwait_wait_event(&sem->writer, readers_active_check(sem), TASK_UNINTERRUPTIBLE);
+	trace_contention_end(sem, 0);
 }
 EXPORT_SYMBOL_GPL(percpu_down_write);
 

kernel/locking/qrwlock.c

@@ -12,6 +12,7 @@
 #include <linux/percpu.h>
 #include <linux/hardirq.h>
 #include <linux/spinlock.h>
+#include <trace/events/lock.h>
 
 /**
  * queued_read_lock_slowpath - acquire read lock of a queue rwlock
@@ -34,6 +35,8 @@ void queued_read_lock_slowpath(struct qrwlock *lock)
 	}
 	atomic_sub(_QR_BIAS, &lock->cnts);
 
+	trace_contention_begin(lock, LCB_F_SPIN | LCB_F_READ);
+
 	/*
 	 * Put the reader into the wait queue
 	 */
@@ -51,6 +54,8 @@ void queued_read_lock_slowpath(struct qrwlock *lock)
 	 * Signal the next one in queue to become queue head
 	 */
 	arch_spin_unlock(&lock->wait_lock);
+
+	trace_contention_end(lock, 0);
 }
 EXPORT_SYMBOL(queued_read_lock_slowpath);
 
@@ -62,6 +67,8 @@ void queued_write_lock_slowpath(struct qrwlock *lock)
 {
 	int cnts;
 
+	trace_contention_begin(lock, LCB_F_SPIN | LCB_F_WRITE);
+
 	/* Put the writer into the wait queue */
 	arch_spin_lock(&lock->wait_lock);
 
@@ -79,5 +86,7 @@ void queued_write_lock_slowpath(struct qrwlock *lock)
 	} while (!atomic_try_cmpxchg_acquire(&lock->cnts, &cnts, _QW_LOCKED));
 unlock:
 	arch_spin_unlock(&lock->wait_lock);
+
+	trace_contention_end(lock, 0);
 }
 EXPORT_SYMBOL(queued_write_lock_slowpath);

kernel/locking/qspinlock.c

@@ -22,6 +22,7 @@
 #include <linux/prefetch.h>
 #include <asm/byteorder.h>
 #include <asm/qspinlock.h>
+#include <trace/events/lock.h>
 
 /*
  * Include queued spinlock statistics code
@@ -401,6 +402,8 @@ void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
 	idx = node->count++;
 	tail = encode_tail(smp_processor_id(), idx);
 
+	trace_contention_begin(lock, LCB_F_SPIN);
+
 	/*
 	 * 4 nodes are allocated based on the assumption that there will
 	 * not be nested NMIs taking spinlocks. That may not be true in
@@ -554,6 +557,8 @@ void queued_spin_lock_slowpath(struct qspinlock *lock, u32 val)
 	pv_kick_node(lock, next);
 
 release:
+	trace_contention_end(lock, 0);
+
 	/*
 	 * release the node
 	 */

kernel/locking/rtmutex.c

@@ -24,6 +24,8 @@
 #include <linux/sched/wake_q.h>
 #include <linux/ww_mutex.h>
 
+#include <trace/events/lock.h>
+
 #include "rtmutex_common.h"
 
 #ifndef WW_RT
@@ -1579,6 +1581,8 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock,
 
 	set_current_state(state);
 
+	trace_contention_begin(lock, LCB_F_RT);
+
 	ret = task_blocks_on_rt_mutex(lock, waiter, current, ww_ctx, chwalk);
 	if (likely(!ret))
 		ret = rt_mutex_slowlock_block(lock, ww_ctx, state, NULL, waiter);
@@ -1601,6 +1605,9 @@ static int __sched __rt_mutex_slowlock(struct rt_mutex_base *lock,
 	 * unconditionally. We might have to fix that up.
 	 */
 	fixup_rt_mutex_waiters(lock);
+
+	trace_contention_end(lock, ret);
+
 	return ret;
 }
 
@@ -1683,6 +1690,8 @@ static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock)
 	/* Save current state and set state to TASK_RTLOCK_WAIT */
 	current_save_and_set_rtlock_wait_state();
 
+	trace_contention_begin(lock, LCB_F_RT);
+
 	task_blocks_on_rt_mutex(lock, &waiter, current, NULL, RT_MUTEX_MIN_CHAINWALK);
 
 	for (;;) {
@@ -1712,6 +1721,8 @@ static void __sched rtlock_slowlock_locked(struct rt_mutex_base *lock)
 	 */
 	fixup_rt_mutex_waiters(lock);
 	debug_rt_mutex_free_waiter(&waiter);
+
+	trace_contention_end(lock, 0);
 }
 
 static __always_inline void __sched rtlock_slowlock(struct rt_mutex_base *lock)

kernel/locking/rwbase_rt.c

@@ -112,6 +112,8 @@ static int __sched __rwbase_read_lock(struct rwbase_rt *rwb,
 	 * Reader2 to call up_read(), which might be unbound.
 	 */
 
+	trace_contention_begin(rwb, LCB_F_RT | LCB_F_READ);
+
 	/*
 	 * For rwlocks this returns 0 unconditionally, so the below
 	 * !ret conditionals are optimized out.
@@ -130,6 +132,8 @@ static int __sched __rwbase_read_lock(struct rwbase_rt *rwb,
 	raw_spin_unlock_irq(&rtm->wait_lock);
 	if (!ret)
 		rwbase_rtmutex_unlock(rtm);
+
+	trace_contention_end(rwb, ret);
 	return ret;
 }
 
@@ -247,11 +251,13 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
 		goto out_unlock;
 
 	rwbase_set_and_save_current_state(state);
+	trace_contention_begin(rwb, LCB_F_RT | LCB_F_WRITE);
 	for (;;) {
 		/* Optimized out for rwlocks */
 		if (rwbase_signal_pending_state(state, current)) {
 			rwbase_restore_current_state();
 			__rwbase_write_unlock(rwb, 0, flags);
+			trace_contention_end(rwb, -EINTR);
 			return -EINTR;
 		}
 
@@ -265,6 +271,7 @@ static int __sched rwbase_write_lock(struct rwbase_rt *rwb,
 		set_current_state(state);
 	}
 	rwbase_restore_current_state();
+	trace_contention_end(rwb, 0);
 
 out_unlock:
 	raw_spin_unlock_irqrestore(&rtm->wait_lock, flags);

kernel/locking/rwsem.c

@@ -27,6 +27,7 @@
 #include <linux/export.h>
 #include <linux/rwsem.h>
 #include <linux/atomic.h>
+#include <trace/events/lock.h>
 
 #ifndef CONFIG_PREEMPT_RT
 #include "lock_events.h"
@@ -1056,6 +1057,8 @@ rwsem_down_read_slowpath(struct rw_semaphore *sem, long count, unsigned int stat
 	if (!wake_q_empty(&wake_q))
 		wake_up_q(&wake_q);
 
+	trace_contention_begin(sem, LCB_F_READ);
+
 	/* wait to be given the lock */
 	for (;;) {
 		set_current_state(state);
@@ -1077,12 +1080,14 @@ rwsem_down_read_slowpath(struct rw_semaphore *sem, long count, unsigned int stat
 
 	__set_current_state(TASK_RUNNING);
 	lockevent_inc(rwsem_rlock);
+	trace_contention_end(sem, 0);
 	return sem;
 
 out_nolock:
 	rwsem_del_wake_waiter(sem, &waiter, &wake_q);
 	__set_current_state(TASK_RUNNING);
 	lockevent_inc(rwsem_rlock_fail);
+	trace_contention_end(sem, -EINTR);
 	return ERR_PTR(-EINTR);
 }
 
@@ -1132,6 +1137,8 @@ rwsem_down_write_slowpath(struct rw_semaphore *sem, int state)
 
 	/* wait until we successfully acquire the lock */
 	set_current_state(state);
+	trace_contention_begin(sem, LCB_F_WRITE);
+
 	for (;;) {
 		if (rwsem_try_write_lock(sem, &waiter)) {
 			/* rwsem_try_write_lock() implies ACQUIRE on success */
@@ -1171,6 +1178,7 @@ rwsem_down_write_slowpath(struct rw_semaphore *sem, int state)
 	__set_current_state(TASK_RUNNING);
 	raw_spin_unlock_irq(&sem->wait_lock);
 	lockevent_inc(rwsem_wlock);
+	trace_contention_end(sem, 0);
 	return sem;
 
 out_nolock:
@@ -1178,6 +1186,7 @@ rwsem_down_write_slowpath(struct rw_semaphore *sem, int state)
 	raw_spin_lock_irq(&sem->wait_lock);
 	rwsem_del_wake_waiter(sem, &waiter, &wake_q);
 	lockevent_inc(rwsem_wlock_fail);
+	trace_contention_end(sem, -EINTR);
 	return ERR_PTR(-EINTR);
 }
 

kernel/locking/semaphore.c

@@ -32,6 +32,7 @@
 #include <linux/semaphore.h>
 #include <linux/spinlock.h>
 #include <linux/ftrace.h>
+#include <trace/events/lock.h>
 
 static noinline void __down(struct semaphore *sem);
 static noinline int __down_interruptible(struct semaphore *sem);
@@ -205,7 +206,7 @@ struct semaphore_waiter {
  * constant, and thus optimised away by the compiler.  Likewise the
  * 'timeout' parameter for the cases without timeouts.
  */
-static inline int __sched __down_common(struct semaphore *sem, long state,
+static inline int __sched ___down_common(struct semaphore *sem, long state,
 								long timeout)
 {
 	struct semaphore_waiter waiter;
@@ -236,6 +237,18 @@ static inline int __sched __down_common(struct semaphore *sem, long state,
 	return -EINTR;
 }
 
+static inline int __sched __down_common(struct semaphore *sem, long state,
+					long timeout)
+{
+	int ret;
+
+	trace_contention_begin(sem, 0);
+	ret = ___down_common(sem, state, timeout);
+	trace_contention_end(sem, ret);
+
+	return ret;
+}
+
 static noinline void __sched __down(struct semaphore *sem)
 {
 	__down_common(sem, TASK_UNINTERRUPTIBLE, MAX_SCHEDULE_TIMEOUT);