ftrace: add tracepoint for timer

new file mode 100644
index 0000000..a664afa
@@ -0,0 +1,134 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM timer
+
+#if !defined(_TRACE_TIMER_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_TIMER_H
+
+#include <linux/tracepoint.h>
+#include <linux/timer.h>
+
+/**
+ * timer_init - called when the timer is initialized
+ * @timer: pointer to struct timer_list
+ **/
+TRACE_EVENT(timer_init,
+
+	TP_PROTO(struct timer_list *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+	),
+
+	TP_printk("timer=%p", __entry->timer)
+);
+
+/**
+ * timer_start - called when the timer is started
+ * @timer: pointer to struct timer_list
+ * @expires: the timer's expire time
+ **/
+TRACE_EVENT(timer_start,
+
+	TP_PROTO(struct timer_list *timer, unsigned long expires),
+
+	TP_ARGS(timer, expires),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer		)
+		__field( void *,	function	)
+		__field( unsigned long,	expires		)
+		__field( int,		timeout		)
+	),
+
+	TP_fast_assign(
+		__entry->timer		= timer;
+		__entry->function	= timer->function;
+		__entry->expires	= expires;
+		__entry->timeout	= timer->expires - jiffies;
+	),
+
+	TP_printk("timer=%p func=%pf expires=%lu timeout=%d", __entry->timer,
+		  __entry->function, __entry->expires, __entry->timeout)
+);
+
+/**
+ * timer_expire - called immediately before the timer callback
+ * @timer: pointer to struct timer_list
+ *
+ * When used in combination with the timer_callback_done tracepoint we can
+ * determine the timer latency.
+ **/
+TRACE_EVENT(timer_expire,
+
+	TP_PROTO(struct timer_list *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+		__field( unsigned long,	jiffies	)
+	),
+
+	TP_fast_assign(
+		__entry->timer		= timer;
+		__entry->jiffies	= jiffies;
+	),
+
+	TP_printk("timer=%p jiffies=%lu", __entry->timer, __entry->jiffies)
+);
+
+/**
+ * timer_callback_done - called immediately after the timer callback returns
+ * @timer: pointer to struct timer_list
+ *
+ * When used in combination with the timer_expire tracepoint we can
+ * determine the timer latency.
+ **/
+TRACE_EVENT(timer_callback_done,
+
+	TP_PROTO(struct timer_list *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field(void *,	timer	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+	),
+
+	TP_printk("func=%p", __entry->timer)
+);
+
+/**
+ * timer_cancel - called when the timer is canceled
+ * @timer: pointer to struct timer_list
+ **/
+TRACE_EVENT(timer_cancel,
+
+	TP_PROTO(struct timer_list *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+	),
+
+	TP_printk("timer=%p", __entry->timer)
+);
+
+#endif /*  _TRACE_TIMER_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>

index 33fc9d1..7b60ff6 100644
@@ -46,6 +46,9 @@
 #include <asm/timex.h>
 #include <asm/io.h>
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/timer.h>
+
 u64 jiffies_64 __cacheline_aligned_in_smp = INITIAL_JIFFIES;
 
 EXPORT_SYMBOL(jiffies_64);

@@ -521,6 +524,25 @@ static inline void debug_timer_activate(struct timer_list *timer) { }
 static inline void debug_timer_deactivate(struct timer_list *timer) { }
 #endif
 
+static inline void debug_init(struct timer_list *timer)
+{
+	debug_timer_init(timer);
+	trace_timer_init(timer);
+}
+
+static inline void
+debug_activate(struct timer_list *timer, unsigned long expires)
+{
+	debug_timer_activate(timer);
+	trace_timer_start(timer, expires);
+}
+
+static inline void debug_deactivate(struct timer_list *timer)
+{
+	debug_timer_deactivate(timer);
+	trace_timer_cancel(timer);
+}
+
 static void __init_timer(struct timer_list *timer,
 			 const char *name,
 			 struct lock_class_key *key)

@@ -549,7 +571,7 @@ void init_timer_key(struct timer_list *timer,
 		    const char *name,
 		    struct lock_class_key *key)
 {
-	debug_timer_init(timer);
+	debug_init(timer);
 	__init_timer(timer, name, key);
 }
 EXPORT_SYMBOL(init_timer_key);

@@ -568,7 +590,7 @@ static inline void detach_timer(struct timer_list *timer,
 {
 	struct list_head *entry = &timer->entry;
 
-	debug_timer_deactivate(timer);
+	debug_deactivate(timer);
 
 	__list_del(entry->prev, entry->next);
 	if (clear_pending)

@@ -632,7 +654,7 @@ __mod_timer(struct timer_list *timer, unsigned long expires,
 			goto out_unlock;
 	}
 
-	debug_timer_activate(timer);
+	debug_activate(timer, expires);
 
 	new_base = __get_cpu_var(tvec_bases);
 

@@ -787,7 +809,7 @@ void add_timer_on(struct timer_list *timer, int cpu)
 	BUG_ON(timer_pending(timer) || !timer->function);
 	spin_lock_irqsave(&base->lock, flags);
 	timer_set_base(timer, base);
-	debug_timer_activate(timer);
+	debug_activate(timer, timer->expires);
 	if (time_before(timer->expires, base->next_timer) &&
 	    !tbase_get_deferrable(timer->base))
 		base->next_timer = timer->expires;

@@ -1000,7 +1022,9 @@ static inline void __run_timers(struct tvec_base *base)
 				 */
 				lock_map_acquire(&lockdep_map);
 
+				trace_timer_expire(timer);
 				fn(data);
+				trace_timer_callback_done(timer);
 
 				lock_map_release(&lockdep_map);
 

index a664afa..dd2f133 100644
@@ -128,6 +128,139 @@ TRACE_EVENT(timer_cancel,
 	TP_printk("timer=%p", __entry->timer)
 );
 
+/**
+ * hrtimer_init - called when the hrtimer is initialized
+ * @timer: pointer to struct hrtimer
+ * @clockid: the hrtimer's clock
+ * @mode: the hrtimer's mode
+ **/
+TRACE_EVENT(hrtimer_init,
+
+	TP_PROTO(struct hrtimer *timer, clockid_t clockid, enum hrtimer_mode mode),
+
+	TP_ARGS(timer, clockid, mode),
+
+	TP_STRUCT__entry(
+		__field( void *,		timer		)
+		__field( clockid_t,		clockid		)
+		__field( enum hrtimer_mode,	mode		)
+	),
+
+	TP_fast_assign(
+		__entry->timer		= timer;
+		__entry->clockid	= clockid;
+		__entry->mode		= mode;
+	),
+
+	TP_printk("timer=%p clockid=%s mode=%s", __entry->timer,
+		  __entry->clockid == CLOCK_REALTIME ?
+			"CLOCK_REALTIME" : "CLOCK_MONOTONIC",
+		  __entry->mode == HRTIMER_MODE_ABS ?
+			"HRTIMER_MODE_ABS" : "HRTIMER_MODE_REL")
+);
+
+/**
+ * hrtimer_start - called when the hrtimer is started
+ * @timer: pointer to struct hrtimer
+ **/
+TRACE_EVENT(hrtimer_start,
+
+	TP_PROTO(struct hrtimer *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer		)
+		__field( void *,	function	)
+		__field( s64,		expires		)
+		__field( s64,		softexpires	)
+	),
+
+	TP_fast_assign(
+		__entry->timer		= timer;
+		__entry->function	= timer->function;
+		__entry->expires	= hrtimer_get_expires(timer).tv64;
+		__entry->softexpires	= hrtimer_get_softexpires(timer).tv64;
+	),
+
+	TP_printk("timer=%p func=%pf expires=%lluns softexpires=%lluns",
+		  __entry->timer, __entry->function,
+		 (unsigned long long)__entry->expires,
+		 (unsigned long long)__entry->softexpires)
+);
+
+/**
+ * htimmer_expire - called immediately before the hrtimer callback
+ * @timer: pointer to struct hrtimer
+ *
+ * When used in combination with the hrtimer_callback_done tracepoint we can
+ * determine the high res timer latency.
+ **/
+TRACE_EVENT(hrtimer_expire,
+
+	TP_PROTO(struct hrtimer *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+		__field( s64,		now	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+		__entry->now	= timer->base->get_time().tv64;
+	),
+
+	TP_printk("timer=%p now=%lluns", __entry->timer,
+		 (unsigned long long)__entry->now)
+ );
+
+ /**
+  * hrtimer_callback_done - called immediately after the hrtimer callback returns
+  * @timer: pointer to struct hrtimer
+  *
+  * When used in combination with the hrtimer_expire tracepoint we can
+  * determine the high res timer latency.
+  **/
+TRACE_EVENT(hrtimer_callback_done,
+
+	TP_PROTO(struct hrtimer *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+	),
+
+	TP_printk("timer=%p", __entry->timer)
+);
+
+/**
+ * hrtimer_cancel - called when the hrtimer is canceled
+ * @timer: pointer to struct hrtimer
+ **/
+TRACE_EVENT(hrtimer_cancel,
+
+	TP_PROTO(struct hrtimer *timer),
+
+	TP_ARGS(timer),
+
+	TP_STRUCT__entry(
+		__field( void *,	timer	)
+	),
+
+	TP_fast_assign(
+		__entry->timer	= timer;
+	),
+
+	TP_printk("timer=%p", __entry->timer)
+);
+
 #endif /*  _TRACE_TIMER_H */
 
 /* This part must be outside protection */

index e2f91ec..f05c7a1 100644
@@ -48,6 +48,8 @@
 
 #include <asm/uaccess.h>
 
+#include <trace/events/timer.h>
+
 /*
  * The timer bases:
  *

@@ -441,6 +443,26 @@ static inline void debug_hrtimer_activate(struct hrtimer *timer) { }
 static inline void debug_hrtimer_deactivate(struct hrtimer *timer) { }
 #endif
 
+static inline void
+debug_init(struct hrtimer *timer, clockid_t clockid,
+	   enum hrtimer_mode mode)
+{
+	debug_hrtimer_init(timer);
+	trace_hrtimer_init(timer, clockid, mode);
+}
+
+static inline void debug_activate(struct hrtimer *timer)
+{
+	debug_hrtimer_activate(timer);
+	trace_hrtimer_start(timer);
+}
+
+static inline void debug_deactivate(struct hrtimer *timer)
+{
+	debug_hrtimer_deactivate(timer);
+	trace_hrtimer_cancel(timer);
+}
+
 /* High resolution timer related functions */
 #ifdef CONFIG_HIGH_RES_TIMERS
 

@@ -797,7 +819,7 @@ static int enqueue_hrtimer(struct hrtimer *timer,
 	struct hrtimer *entry;
 	int leftmost = 1;
 
-	debug_hrtimer_activate(timer);
+	debug_activate(timer);
 
 	/*
 	 * Find the right place in the rbtree:

@@ -883,7 +905,7 @@ remove_hrtimer(struct hrtimer *timer, struct hrtimer_clock_base *base)
 		 * reprogramming happens in the interrupt handler. This is a
 		 * rare case and less expensive than a smp call.
 		 */
-		debug_hrtimer_deactivate(timer);
+		debug_deactivate(timer);
 		timer_stats_hrtimer_clear_start_info(timer);
 		reprogram = base->cpu_base == &__get_cpu_var(hrtimer_bases);
 		__remove_hrtimer(timer, base, HRTIMER_STATE_INACTIVE,

@@ -1116,7 +1138,7 @@ static void __hrtimer_init(struct hrtimer *timer, clockid_t clock_id,
 void hrtimer_init(struct hrtimer *timer, clockid_t clock_id,
 		  enum hrtimer_mode mode)
 {
-	debug_hrtimer_init(timer);
+	debug_init(timer, clock_id, mode);
 	__hrtimer_init(timer, clock_id, mode);
 }
 EXPORT_SYMBOL_GPL(hrtimer_init);

@@ -1149,7 +1171,7 @@ static void __run_hrtimer(struct hrtimer *timer)
 
 	WARN_ON(!irqs_disabled());
 
-	debug_hrtimer_deactivate(timer);
+	debug_deactivate(timer);
 	__remove_hrtimer(timer, base, HRTIMER_STATE_CALLBACK, 0);
 	timer_stats_account_hrtimer(timer);
 	fn = timer->function;

@@ -1160,7 +1182,9 @@ static void __run_hrtimer(struct hrtimer *timer)
 	 * the timer base.
 	 */
 	spin_unlock(&cpu_base->lock);
+	trace_hrtimer_expire(timer);
 	restart = fn(timer);
+	trace_hrtimer_callback_done(timer);
 	spin_lock(&cpu_base->lock);
 
 	/*

@@ -1569,7 +1593,7 @@ static void migrate_hrtimer_list(struct hrtimer_clock_base *old_base,
 	while ((node = rb_first(&old_base->active))) {
 		timer = rb_entry(node, struct hrtimer, node);
 		BUG_ON(hrtimer_callback_running(timer));
-		debug_hrtimer_deactivate(timer);
+		debug_deactivate(timer);
 
 		/*
 		 * Mark it as STATE_MIGRATE not INACTIVE otherwise the

index dd2f133..11bdb47 100644
@@ -261,6 +261,71 @@ TRACE_EVENT(hrtimer_cancel,
 	TP_printk("timer=%p", __entry->timer)
 );
 
+/**
+ * itimer_state - called when itimer is started or canceled
+ * @which: name of the interval timer
+ * @value: the itimer's value, itimer is canceled if value->it_value is zero,
+ * 	   otherwise is started
+ * @expires: the itimer's expire time
+ **/
+TRACE_EVENT(itimer_state,
+
+	TP_PROTO(int which, const struct itimerval *const value, cputime_t expires),
+
+	TP_ARGS(which, value, expires),
+
+	TP_STRUCT__entry(
+		__field(	int,		which		)
+		__field(	cputime_t,	expires		)
+		__field(	long,		value_sec	)
+		__field(	long,		value_usec	)
+		__field(	long,		interval_sec	)
+		__field(	long,		interval_usec	)
+	),
+
+	TP_fast_assign(
+		__entry->which		= which;
+		__entry->expires	= expires;
+		__entry->value_sec	= value->it_value.tv_sec;
+		__entry->value_usec	= value->it_value.tv_usec;
+		__entry->interval_sec	= value->it_interval.tv_sec;
+		__entry->interval_usec	= value->it_interval.tv_usec;
+	),
+
+	TP_printk("which=%d expires=%lu it_value=%lu.%lu it_interval=%lu.%lu",
+		  __entry->which, __entry->expires,
+		  __entry->value_sec, __entry->value_usec,
+		  __entry->interval_sec, __entry->interval_usec)
+);
+
+/**
+ * itimer_expire: called when itimer expires
+ * @which: name of the interval timer
+ * @pid: pid of process which owns the itimer
+ * @now: current time, used to calculate the latency of itimer
+ **/
+TRACE_EVENT(itimer_expire,
+
+	TP_PROTO(int which, struct pid *pid, cputime_t now),
+
+	TP_ARGS(which, pid, now),
+
+	TP_STRUCT__entry(
+		__field( int ,		which	)
+		__field( int,		pid	)
+		__field( cputime_t,	now	)
+	),
+
+	TP_fast_assign(
+		__entry->which	= which;
+		__entry->now	= now;
+		__entry->pid	= pid_vnr(pid);
+	),
+
+	TP_printk("pid=%d which=%d now=%lu", __entry->pid, __entry->which,
+		  __entry->now)
+);
+
 #endif /*  _TRACE_TIMER_H */
 
 /* This part must be outside protection */

index 8078a32..a3ceb34 100644
@@ -12,6 +12,7 @@
 #include <linux/time.h>
 #include <linux/posix-timers.h>
 #include <linux/hrtimer.h>
+#include <trace/events/timer.h>
 
 #include <asm/uaccess.h>
 

@@ -122,6 +123,7 @@ enum hrtimer_restart it_real_fn(struct hrtimer *timer)
 	struct signal_struct *sig =
 		container_of(timer, struct signal_struct, real_timer);
 
+	trace_itimer_expire(ITIMER_REAL, sig->leader_pid, 0);
 	kill_pid_info(SIGALRM, SEND_SIG_PRIV, sig->leader_pid);
 
 	return HRTIMER_NORESTART;

@@ -166,6 +168,8 @@ static void set_cpu_itimer(struct task_struct *tsk, unsigned int clock_id,
 	}
 	it->expires = nval;
 	it->incr = ninterval;
+	trace_itimer_state(clock_id == CPUCLOCK_VIRT ? ITIMER_VIRTUAL : ITIMER_PROF,
+			   value, nval);
 
 	spin_unlock_irq(&tsk->sighand->siglock);
 

@@ -217,6 +221,7 @@ again:
 		} else
 			tsk->signal->it_real_incr.tv64 = 0;
 
+		trace_itimer_state(ITIMER_REAL, value, 0);
 		spin_unlock_irq(&tsk->sighand->siglock);
 		break;
 	case ITIMER_VIRTUAL:

index 18bdde6..f1dbab9 100644
@@ -8,6 +8,7 @@
 #include <linux/math64.h>
 #include <asm/uaccess.h>
 #include <linux/kernel_stat.h>
+#include <trace/events/timer.h>
 
 /*
  * Called after updating RLIMIT_CPU to set timer expiration if necessary.

@@ -1092,6 +1093,8 @@ static void check_cpu_itimer(struct task_struct *tsk, struct cpu_itimer *it,
 		} else
 			it->expires = cputime_zero;
 
+		trace_itimer_expire(signo == SIGPROF ? ITIMER_PROF : ITIMER_VIRTUAL,
+				    tsk->signal->leader_pid, cur_time);
 		__group_send_sig_info(signo, SEND_SIG_PRIV, tsk);
 	}