[linux-2.6-block.git] / include / linux / tracepoint.h

#ifndef _LINUX_TRACEPOINT_H
#define _LINUX_TRACEPOINT_H

/*
 * Kernel Tracepoint API.
 *
 * See Documentation/trace/tracepoints.txt.
 *
 * (C) Copyright 2008 Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
 *
 * Heavily inspired from the Linux Kernel Markers.
 *
 * This file is released under the GPLv2.
 * See the file COPYING for more details.
 */

#include <linux/types.h>
#include <linux/rcupdate.h>

struct module;
struct tracepoint;

struct tracepoint {
	const char *name;		/* Tracepoint name */
	int state;			/* State. */
	void (*regfunc)(void);
	void (*unregfunc)(void);
	void **funcs;
} __attribute__((aligned(32)));		/*
					 * Aligned on 32 bytes because it is
					 * globally visible and gcc happily
					 * align these on the structure size.
					 * Keep in sync with vmlinux.lds.h.
					 */

#ifndef DECLARE_TRACE

#define TP_PROTO(args...)	args
#define TP_ARGS(args...)	args

#ifdef CONFIG_TRACEPOINTS

/*
 * it_func[0] is never NULL because there is at least one element in the array
 * when the array itself is non NULL.
 */
#define __DO_TRACE(tp, proto, args)					\
	do {								\
		void **it_func;						\
									\
		rcu_read_lock_sched_notrace();				\
		it_func = rcu_dereference_sched((tp)->funcs);		\
		if (it_func) {						\
			do {						\
				((void(*)(proto))(*it_func))(args);	\
			} while (*(++it_func));				\
		}							\
		rcu_read_unlock_sched_notrace();			\
	} while (0)

/*
 * Make sure the alignment of the structure in the __tracepoints section will
 * not add unwanted padding between the beginning of the section and the
 * structure. Force alignment to the same alignment as the section start.
 */
#define DECLARE_TRACE(name, proto, args)				\
	extern struct tracepoint __tracepoint_##name;			\
	static inline void trace_##name(proto)				\
	{								\
		if (unlikely(__tracepoint_##name.state))		\
			__DO_TRACE(&__tracepoint_##name,		\
				TP_PROTO(proto), TP_ARGS(args));	\
	}								\
	static inline int register_trace_##name(void (*probe)(proto))	\
	{								\
		return tracepoint_probe_register(#name, (void *)probe);	\
	}								\
	static inline int unregister_trace_##name(void (*probe)(proto))	\
	{								\
		return tracepoint_probe_unregister(#name, (void *)probe);\
	}


#define DEFINE_TRACE_FN(name, reg, unreg)				\
	static const char __tpstrtab_##name[]				\
	__attribute__((section("__tracepoints_strings"))) = #name;	\
	struct tracepoint __tracepoint_##name				\
	__attribute__((section("__tracepoints"), aligned(32))) =	\
		{ __tpstrtab_##name, 0, reg, unreg, NULL }

#define DEFINE_TRACE(name)						\
	DEFINE_TRACE_FN(name, NULL, NULL);

#define EXPORT_TRACEPOINT_SYMBOL_GPL(name)				\
	EXPORT_SYMBOL_GPL(__tracepoint_##name)
#define EXPORT_TRACEPOINT_SYMBOL(name)					\
	EXPORT_SYMBOL(__tracepoint_##name)

extern void tracepoint_update_probe_range(struct tracepoint *begin,
	struct tracepoint *end);

#else /* !CONFIG_TRACEPOINTS */
#define DECLARE_TRACE(name, proto, args)				\
	static inline void _do_trace_##name(struct tracepoint *tp, proto) \
	{ }								\
	static inline void trace_##name(proto)				\
	{ }								\
	static inline int register_trace_##name(void (*probe)(proto))	\
	{								\
		return -ENOSYS;						\
	}								\
	static inline int unregister_trace_##name(void (*probe)(proto))	\
	{								\
		return -ENOSYS;						\
	}

#define DEFINE_TRACE_FN(name, reg, unreg)
#define DEFINE_TRACE(name)
#define EXPORT_TRACEPOINT_SYMBOL_GPL(name)
#define EXPORT_TRACEPOINT_SYMBOL(name)

static inline void tracepoint_update_probe_range(struct tracepoint *begin,
	struct tracepoint *end)
{ }
#endif /* CONFIG_TRACEPOINTS */
#endif /* DECLARE_TRACE */

/*
 * Connect a probe to a tracepoint.
 * Internal API, should not be used directly.
 */
extern int tracepoint_probe_register(const char *name, void *probe);

/*
 * Disconnect a probe from a tracepoint.
 * Internal API, should not be used directly.
 */
extern int tracepoint_probe_unregister(const char *name, void *probe);

extern int tracepoint_probe_register_noupdate(const char *name, void *probe);
extern int tracepoint_probe_unregister_noupdate(const char *name, void *probe);
extern void tracepoint_probe_update_all(void);

struct tracepoint_iter {
	struct module *module;
	struct tracepoint *tracepoint;
};

extern void tracepoint_iter_start(struct tracepoint_iter *iter);
extern void tracepoint_iter_next(struct tracepoint_iter *iter);
extern void tracepoint_iter_stop(struct tracepoint_iter *iter);
extern void tracepoint_iter_reset(struct tracepoint_iter *iter);
extern int tracepoint_get_iter_range(struct tracepoint **tracepoint,
	struct tracepoint *begin, struct tracepoint *end);

/*
 * tracepoint_synchronize_unregister must be called between the last tracepoint
 * probe unregistration and the end of module exit to make sure there is no
 * caller executing a probe when it is freed.
 */
static inline void tracepoint_synchronize_unregister(void)
{
	synchronize_sched();
}

#define PARAMS(args...) args

#endif /* _LINUX_TRACEPOINT_H */

/*
 * Note: we keep the TRACE_EVENT outside the include file ifdef protection.
 *  This is due to the way trace events work. If a file includes two
 *  trace event headers under one "CREATE_TRACE_POINTS" the first include
 *  will override the TRACE_EVENT and break the second include.
 */

#ifndef TRACE_EVENT
/*
 * For use with the TRACE_EVENT macro:
 *
 * We define a tracepoint, its arguments, its printk format
 * and its 'fast binay record' layout.
 *
 * Firstly, name your tracepoint via TRACE_EVENT(name : the
 * 'subsystem_event' notation is fine.
 *
 * Think about this whole construct as the
 * 'trace_sched_switch() function' from now on.
 *
 *
 *  TRACE_EVENT(sched_switch,
 *
 *	*
 *	* A function has a regular function arguments
 *	* prototype, declare it via TP_PROTO():
 *	*
 *
 *	TP_PROTO(struct rq *rq, struct task_struct *prev,
 *		 struct task_struct *next),
 *
 *	*
 *	* Define the call signature of the 'function'.
 *	* (Design sidenote: we use this instead of a
 *	*  TP_PROTO1/TP_PROTO2/TP_PROTO3 ugliness.)
 *	*
 *
 *	TP_ARGS(rq, prev, next),
 *
 *	*
 *	* Fast binary tracing: define the trace record via
 *	* TP_STRUCT__entry(). You can think about it like a
 *	* regular C structure local variable definition.
 *	*
 *	* This is how the trace record is structured and will
 *	* be saved into the ring buffer. These are the fields
 *	* that will be exposed to user-space in
 *	* /sys/kernel/debug/tracing/events/<*>/format.
 *	*
 *	* The declared 'local variable' is called '__entry'
 *	*
 *	* __field(pid_t, prev_prid) is equivalent to a standard declariton:
 *	*
 *	*	pid_t	prev_pid;
 *	*
 *	* __array(char, prev_comm, TASK_COMM_LEN) is equivalent to:
 *	*
 *	*	char	prev_comm[TASK_COMM_LEN];
 *	*
 *
 *	TP_STRUCT__entry(
 *		__array(	char,	prev_comm,	TASK_COMM_LEN	)
 *		__field(	pid_t,	prev_pid			)
 *		__field(	int,	prev_prio			)
 *		__array(	char,	next_comm,	TASK_COMM_LEN	)
 *		__field(	pid_t,	next_pid			)
 *		__field(	int,	next_prio			)
 *	),
 *
 *	*
 *	* Assign the entry into the trace record, by embedding
 *	* a full C statement block into TP_fast_assign(). You
 *	* can refer to the trace record as '__entry' -
 *	* otherwise you can put arbitrary C code in here.
 *	*
 *	* Note: this C code will execute every time a trace event
 *	* happens, on an active tracepoint.
 *	*
 *
 *	TP_fast_assign(
 *		memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
 *		__entry->prev_pid	= prev->pid;
 *		__entry->prev_prio	= prev->prio;
 *		memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
 *		__entry->next_pid	= next->pid;
 *		__entry->next_prio	= next->prio;
 *	)
 *
 *	*
 *	* Formatted output of a trace record via TP_printk().
 *	* This is how the tracepoint will appear under ftrace
 *	* plugins that make use of this tracepoint.
 *	*
 *	* (raw-binary tracing wont actually perform this step.)
 *	*
 *
 *	TP_printk("task %s:%d [%d] ==> %s:%d [%d]",
 *		__entry->prev_comm, __entry->prev_pid, __entry->prev_prio,
 *		__entry->next_comm, __entry->next_pid, __entry->next_prio),
 *
 * );
 *
 * This macro construct is thus used for the regular printk format
 * tracing setup, it is used to construct a function pointer based
 * tracepoint callback (this is used by programmatic plugins and
 * can also by used by generic instrumentation like SystemTap), and
 * it is also used to expose a structured trace record in
 * /sys/kernel/debug/tracing/events/.
 *
 * A set of (un)registration functions can be passed to the variant
 * TRACE_EVENT_FN to perform any (un)registration work.
 */

#define DECLARE_EVENT_CLASS(name, proto, args, tstruct, assign, print)
#define DEFINE_EVENT(template, name, proto, args)		\
	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
#define DEFINE_EVENT_PRINT(template, name, proto, args, print)	\
	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))

#define TRACE_EVENT(name, proto, args, struct, assign, print)	\
	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
#define TRACE_EVENT_FN(name, proto, args, struct,		\
		assign, print, reg, unreg)			\
	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))

#endif /* ifdef TRACE_EVENT (see note above) */
Commit	Line	Data
97e1c18e MD	1	#ifndef _LINUX_TRACEPOINT_H
	2	#define _LINUX_TRACEPOINT_H
	3
	4	/*
	5	* Kernel Tracepoint API.
	6	*
8cd09a59	7	* See Documentation/trace/tracepoints.txt.
97e1c18e MD	8	*
	9	* (C) Copyright 2008 Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
	10	*
	11	* Heavily inspired from the Linux Kernel Markers.
	12	*
	13	* This file is released under the GPLv2.
	14	* See the file COPYING for more details.
	15	*/
	16
	17	#include <linux/types.h>
	18	#include <linux/rcupdate.h>
	19
	20	struct module;
	21	struct tracepoint;
	22
	23	struct tracepoint {
	24	const char name; / Tracepoint name */
	25	int state; /* State. */
97419875 JS	26	void (*regfunc)(void);
97419875 JS	27	void (*unregfunc)(void);
97e1c18e	28	void **funcs;
7e066fb8 MD	29	} __attribute__((aligned(32))); /*
	30	* Aligned on 32 bytes because it is
	31	* globally visible and gcc happily
	32	* align these on the structure size.
	33	* Keep in sync with vmlinux.lds.h.
	34	*/
97e1c18e	35
ea20d929 SR	36	#ifndef DECLARE_TRACE
ea20d929 SR	37
2939b046	38	#define TP_PROTO(args...) args
8cd09a59	39	#define TP_ARGS(args...) args
97e1c18e MD	40
	41	#ifdef CONFIG_TRACEPOINTS
	42
	43	/*
	44	* it_func[0] is never NULL because there is at least one element in the array
	45	* when the array itself is non NULL.
	46	*/
	47	#define __DO_TRACE(tp, proto, args) \
	48	do { \
	49	void **it_func; \
	50	\
da7b3eab	51	rcu_read_lock_sched_notrace(); \
7f5b7742	52	it_func = rcu_dereference_sched((tp)->funcs); \
97e1c18e MD	53	if (it_func) { \
	54	do { \
	55	((void()(proto))(it_func))(args); \
	56	} while (*(++it_func)); \
	57	} \
da7b3eab	58	rcu_read_unlock_sched_notrace(); \
97e1c18e MD	59	} while (0)
	60
	61	/*
	62	* Make sure the alignment of the structure in the __tracepoints section will
	63	* not add unwanted padding between the beginning of the section and the
	64	* structure. Force alignment to the same alignment as the section start.
	65	*/
97419875	66	#define DECLARE_TRACE(name, proto, args) \
7e066fb8	67	extern struct tracepoint __tracepoint_##name; \
97e1c18e MD	68	static inline void trace_##name(proto) \
97e1c18e MD	69	{ \
97e1c18e MD	70	if (unlikely(__tracepoint_##name.state)) \
97e1c18e MD	71	__DO_TRACE(&__tracepoint_##name, \
2939b046	72	TP_PROTO(proto), TP_ARGS(args)); \
97e1c18e MD	73	} \
	74	static inline int register_trace_##name(void (*probe)(proto)) \
	75	{ \
97419875	76	return tracepoint_probe_register(#name, (void *)probe); \
97e1c18e	77	} \
c420970e	78	static inline int unregister_trace_##name(void (*probe)(proto)) \
97e1c18e	79	{ \
97419875	80	return tracepoint_probe_unregister(#name, (void *)probe);\
97e1c18e MD	81	}
97e1c18e MD	82
63fbdab3	83
97419875	84	#define DEFINE_TRACE_FN(name, reg, unreg) \
7e066fb8 MD	85	static const char __tpstrtab_##name[] \
	86	__attribute__((section("__tracepoints_strings"))) = #name; \
	87	struct tracepoint __tracepoint_##name \
	88	__attribute__((section("__tracepoints"), aligned(32))) = \
97419875 JS	89	{ __tpstrtab_##name, 0, reg, unreg, NULL }
	90
	91	#define DEFINE_TRACE(name) \
	92	DEFINE_TRACE_FN(name, NULL, NULL);
7e066fb8 MD	93
	94	#define EXPORT_TRACEPOINT_SYMBOL_GPL(name) \
	95	EXPORT_SYMBOL_GPL(__tracepoint_##name)
	96	#define EXPORT_TRACEPOINT_SYMBOL(name) \
	97	EXPORT_SYMBOL(__tracepoint_##name)
	98
97e1c18e MD	99	extern void tracepoint_update_probe_range(struct tracepoint *begin,
	100	struct tracepoint *end);
	101
	102	#else /* !CONFIG_TRACEPOINTS */
97419875	103	#define DECLARE_TRACE(name, proto, args) \
97e1c18e MD	104	static inline void _do_trace_##name(struct tracepoint *tp, proto) \
	105	{ } \
	106	static inline void trace_##name(proto) \
	107	{ } \
	108	static inline int register_trace_##name(void (*probe)(proto)) \
	109	{ \
	110	return -ENOSYS; \
	111	} \
c420970e MD	112	static inline int unregister_trace_##name(void (*probe)(proto)) \
	113	{ \
	114	return -ENOSYS; \
	115	}
97e1c18e	116
97419875	117	#define DEFINE_TRACE_FN(name, reg, unreg)
7e066fb8 MD	118	#define DEFINE_TRACE(name)
	119	#define EXPORT_TRACEPOINT_SYMBOL_GPL(name)
	120	#define EXPORT_TRACEPOINT_SYMBOL(name)
	121
97e1c18e MD	122	static inline void tracepoint_update_probe_range(struct tracepoint *begin,
	123	struct tracepoint *end)
	124	{ }
	125	#endif /* CONFIG_TRACEPOINTS */
ea20d929	126	#endif /* DECLARE_TRACE */
97e1c18e MD	127
	128	/*
	129	* Connect a probe to a tracepoint.
	130	* Internal API, should not be used directly.
	131	*/
	132	extern int tracepoint_probe_register(const char name, void probe);
	133
	134	/*
	135	* Disconnect a probe from a tracepoint.
	136	* Internal API, should not be used directly.
	137	*/
	138	extern int tracepoint_probe_unregister(const char name, void probe);
	139
127cafbb LJ	140	extern int tracepoint_probe_register_noupdate(const char name, void probe);
	141	extern int tracepoint_probe_unregister_noupdate(const char name, void probe);
	142	extern void tracepoint_probe_update_all(void);
	143
97e1c18e MD	144	struct tracepoint_iter {
	145	struct module *module;
	146	struct tracepoint *tracepoint;
	147	};
	148
	149	extern void tracepoint_iter_start(struct tracepoint_iter *iter);
	150	extern void tracepoint_iter_next(struct tracepoint_iter *iter);
	151	extern void tracepoint_iter_stop(struct tracepoint_iter *iter);
	152	extern void tracepoint_iter_reset(struct tracepoint_iter *iter);
	153	extern int tracepoint_get_iter_range(struct tracepoint **tracepoint,
	154	struct tracepoint begin, struct tracepoint end);
	155
f2461fc8 MD	156	/*
	157	* tracepoint_synchronize_unregister must be called between the last tracepoint
	158	* probe unregistration and the end of module exit to make sure there is no
	159	* caller executing a probe when it is freed.
	160	*/
231375cc MD	161	static inline void tracepoint_synchronize_unregister(void)
	162	{
	163	synchronize_sched();
	164	}
f2461fc8	165
3cdfdf91	166	#define PARAMS(args...) args
7cb2e3ee SR	167
	168	#endif /* _LINUX_TRACEPOINT_H */
	169
	170	/*
	171	* Note: we keep the TRACE_EVENT outside the include file ifdef protection.
	172	* This is due to the way trace events work. If a file includes two
	173	* trace event headers under one "CREATE_TRACE_POINTS" the first include
	174	* will override the TRACE_EVENT and break the second include.
	175	*/
ea20d929	176
ea20d929	177	#ifndef TRACE_EVENT
823f9124 SR	178	/*
	179	* For use with the TRACE_EVENT macro:
	180	*
	181	* We define a tracepoint, its arguments, its printk format
	182	* and its 'fast binay record' layout.
	183	*
	184	* Firstly, name your tracepoint via TRACE_EVENT(name : the
	185	* 'subsystem_event' notation is fine.
	186	*
	187	* Think about this whole construct as the
	188	* 'trace_sched_switch() function' from now on.
	189	*
	190	*
	191	* TRACE_EVENT(sched_switch,
	192	*
	193	* *
	194	* * A function has a regular function arguments
	195	* * prototype, declare it via TP_PROTO():
	196	* *
	197	*
ef18012b SR	198	* TP_PROTO(struct rq rq, struct task_struct prev,
ef18012b SR	199	* struct task_struct *next),
823f9124 SR	200	*
	201	* *
	202	* * Define the call signature of the 'function'.
	203	* * (Design sidenote: we use this instead of a
	204	* * TP_PROTO1/TP_PROTO2/TP_PROTO3 ugliness.)
	205	* *
	206	*
ef18012b	207	* TP_ARGS(rq, prev, next),
823f9124 SR	208	*
	209	* *
	210	* * Fast binary tracing: define the trace record via
	211	* * TP_STRUCT__entry(). You can think about it like a
	212	* * regular C structure local variable definition.
	213	* *
	214	* * This is how the trace record is structured and will
	215	* * be saved into the ring buffer. These are the fields
	216	* * that will be exposed to user-space in
156f5a78	217	* * /sys/kernel/debug/tracing/events/<*>/format.
823f9124 SR	218	* *
	219	* * The declared 'local variable' is called '__entry'
	220	* *
	221	* * __field(pid_t, prev_prid) is equivalent to a standard declariton:
	222	* *
	223	* * pid_t prev_pid;
	224	* *
	225	* * __array(char, prev_comm, TASK_COMM_LEN) is equivalent to:
	226	* *
	227	* * char prev_comm[TASK_COMM_LEN];
	228	* *
	229	*
	230	* TP_STRUCT__entry(
	231	* __array( char, prev_comm, TASK_COMM_LEN )
	232	* __field( pid_t, prev_pid )
	233	* __field( int, prev_prio )
	234	* __array( char, next_comm, TASK_COMM_LEN )
	235	* __field( pid_t, next_pid )
	236	* __field( int, next_prio )
	237	* ),
	238	*
	239	* *
	240	* * Assign the entry into the trace record, by embedding
	241	* * a full C statement block into TP_fast_assign(). You
	242	* * can refer to the trace record as '__entry' -
	243	* * otherwise you can put arbitrary C code in here.
	244	* *
	245	* * Note: this C code will execute every time a trace event
	246	* * happens, on an active tracepoint.
	247	* *
	248	*
ef18012b SR	249	* TP_fast_assign(
	250	* memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN);
	251	* __entry->prev_pid = prev->pid;
	252	* __entry->prev_prio = prev->prio;
823f9124 SR	253	* memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN);
823f9124 SR	254	* __entry->next_pid = next->pid;
ef18012b	255	* __entry->next_prio = next->prio;
823f9124 SR	256	* )
	257	*
	258	* *
	259	* * Formatted output of a trace record via TP_printk().
	260	* * This is how the tracepoint will appear under ftrace
	261	* * plugins that make use of this tracepoint.
	262	* *
	263	* * (raw-binary tracing wont actually perform this step.)
	264	* *
	265	*
	266	* TP_printk("task %s:%d [%d] ==> %s:%d [%d]",
	267	* __entry->prev_comm, __entry->prev_pid, __entry->prev_prio,
	268	* __entry->next_comm, __entry->next_pid, __entry->next_prio),
	269	*
	270	* );
	271	*
	272	* This macro construct is thus used for the regular printk format
	273	* tracing setup, it is used to construct a function pointer based
	274	* tracepoint callback (this is used by programmatic plugins and
	275	* can also by used by generic instrumentation like SystemTap), and
	276	* it is also used to expose a structured trace record in
156f5a78	277	* /sys/kernel/debug/tracing/events/.
97419875 JS	278	*
	279	* A set of (un)registration functions can be passed to the variant
	280	* TRACE_EVENT_FN to perform any (un)registration work.
823f9124 SR	281	*/
823f9124 SR	282
091ad365	283	#define DECLARE_EVENT_CLASS(name, proto, args, tstruct, assign, print)
ff038f5c SR	284	#define DEFINE_EVENT(template, name, proto, args) \
ff038f5c SR	285	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
e5bc9721 SR	286	#define DEFINE_EVENT_PRINT(template, name, proto, args, print) \
e5bc9721 SR	287	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
ff038f5c	288
30a8fecc	289	#define TRACE_EVENT(name, proto, args, struct, assign, print) \
da4d0302	290	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
97419875 JS	291	#define TRACE_EVENT_FN(name, proto, args, struct, \
	292	assign, print, reg, unreg) \
	293	DECLARE_TRACE(name, PARAMS(proto), PARAMS(args))
7cb2e3ee SR	294
7cb2e3ee SR	295	#endif /* ifdef TRACE_EVENT (see note above) */