[PATCH 0/5] tracing/syscalls: Add filtering support

[PATCH 0/5] tracing/syscalls: Add filtering support

[Li Zefan]

This patchset adds filtering support for syscall events

For syscall enter events, we can do filtering on syscall
parameters, and for exit events, filter on syscall return
value.

For example, we set 'mode == 0666' (0666 == 0x1b6) to
sys_enter_open, and set 'ret == 0' to sys_exit_open,
and here's the output:

 # echo 'mode == 0666' > events/syscalls/sys_enter_open
 # echo 'ret == 0' > events/syscalls/sys_exit_open
 # echo 1 > events/syscalls/sys_enter_open
 # echo 1 > events/syscalls/sys_exit_open
 # cat trace
 ...
   modprobe-3084 [001] 117.463140: sys_open(filename: 917d3e8, flags: 0, mode: 1b6) 
   modprobe-3084 [001] 117.463176: sys_open -> 0x0
       less-3086 [001] 117.510455: sys_open(filename: 9c6bdb8, flags: 8000, mode: 1b6) 
   sendmail-2574 [001] 122.145840: sys_open(filename: b807a365, flags: 0, mode: 1b6) 
 ...

[PATCH 1/5] tracing/syscalls: Fix fields format for enter events
[PATCH 2/5] tracing/syscalls: Add fields format for exit events
[PATCH 3/5] tracing/events: Add ftrace_event_call param to define_fields()
[PATCH 4/5] tracing/events: Add trace_define_common_fields()
[PATCH 5/5] tracing/syscalls: Add filtering support
---
 include/linux/ftrace_event.h  |   15 ++----
 include/linux/syscalls.h      |   19 +++++--
 include/trace/ftrace.h        |   11 ++---
 include/trace/syscall.h       |   13 ++++-
 kernel/trace/trace_events.c   |   29 +++++++++-
 kernel/trace/trace_export.c   |   13 ++---
 kernel/trace/trace_syscalls.c |  120 +++++++++++++++++++++++++++++++----------
 7 files changed, 156 insertions(+), 64 deletions(-)

[PATCH 1/5] tracing/syscalls: Fix fields format for enter events

[Li Zefan]

The "format" file of a trace event is originally for parsers to
parse ftrace binary output.

But the "format" file of a syscall event can only be used by
perfcounter, because it describes the format of struct
syscall_enter_record not struct syscall_trace_enter.

To fix this, we remove struct syscall_enter_record, and then
struct syscall_trace_enter will be used by both perf profile
and ftrace.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 kernel/trace/trace_syscalls.c |   51 ++++++++++++++++++++++------------------
 1 files changed, 28 insertions(+), 23 deletions(-)

diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index f130dac..d10daf0 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -90,26 +90,39 @@ print_syscall_exit(struct trace_iterator *iter, int flags)
 	return TRACE_TYPE_HANDLED;
 }
 
+extern char *__bad_type_size(void);
+
+#define SYSCALL_FIELD(type, name)					\
+	sizeof(type) != sizeof(trace.name) ?				\
+		__bad_type_size() :					\
+		#type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
+
 int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 {
 	int i;
 	int nr;
-	int ret = 0;
+	int ret;
 	struct syscall_metadata *entry;
-	int offset = sizeof(struct trace_entry);
+	struct syscall_trace_enter trace;
+	int offset = offsetof(struct syscall_trace_enter, args);
 
-	nr = syscall_name_to_nr((char *)call->data);
+	nr = syscall_name_to_nr(call->data);
 	entry = syscall_nr_to_meta(nr);
 
 	if (!entry)
-		return ret;
+		return 0;
+
+	ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+			       SYSCALL_FIELD(int, nr));
+	if (!ret)
+		return 0;
 
 	for (i = 0; i < entry->nb_args; i++) {
 		ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i],
 				        entry->args[i]);
 		if (!ret)
 			return 0;
-		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%lu;\n", offset,
+		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;\n", offset,
 				       sizeof(unsigned long));
 		if (!ret)
 			return 0;
@@ -118,7 +131,7 @@ int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 
 	trace_seq_printf(s, "\nprint fmt: \"");
 	for (i = 0; i < entry->nb_args; i++) {
-		ret = trace_seq_printf(s, "%s: 0x%%0%lulx%s", entry->args[i],
+		ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i],
 				        sizeof(unsigned long),
 					i == entry->nb_args - 1 ? "\", " : ", ");
 		if (!ret)
@@ -287,16 +300,6 @@ struct trace_event event_syscall_exit = {
 
 #ifdef CONFIG_EVENT_PROFILE
 
-struct syscall_enter_record {
-	struct trace_entry	entry;
-	unsigned long		args[0];
-};
-
-struct syscall_exit_record {
-	struct trace_entry	entry;
-	unsigned long		ret;
-};
-
 static DECLARE_BITMAP(enabled_prof_enter_syscalls, FTRACE_SYSCALL_MAX);
 static DECLARE_BITMAP(enabled_prof_exit_syscalls, FTRACE_SYSCALL_MAX);
 static int sys_prof_refcount_enter;
@@ -304,7 +307,7 @@ static int sys_prof_refcount_exit;
 
 static void prof_syscall_enter(struct pt_regs *regs, long id)
 {
-	struct syscall_enter_record *rec;
+	struct syscall_trace_enter *rec;
 	struct syscall_metadata *sys_data;
 	int syscall_nr;
 	int size;
@@ -328,9 +331,10 @@ static void prof_syscall_enter(struct pt_regs *regs, long id)
 		/* zero the dead bytes from align to not leak stack to user */
 		*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
 
-		rec = (struct syscall_enter_record *) raw_data;
-		tracing_generic_entry_update(&rec->entry, 0, 0);
-		rec->entry.type = sys_data->enter_id;
+		rec = (struct syscall_trace_enter *) raw_data;
+		tracing_generic_entry_update(&rec->ent, 0, 0);
+		rec->ent.type = sys_data->enter_id;
+		rec->nr = syscall_nr;
 		syscall_get_arguments(current, regs, 0, sys_data->nb_args,
 				       (unsigned long *)&rec->args);
 		perf_tpcounter_event(sys_data->enter_id, 0, 1, rec, size);
@@ -379,7 +383,7 @@ void unreg_prof_syscall_enter(char *name)
 static void prof_syscall_exit(struct pt_regs *regs, long ret)
 {
 	struct syscall_metadata *sys_data;
-	struct syscall_exit_record rec;
+	struct syscall_trace_exit rec;
 	int syscall_nr;
 
 	syscall_nr = syscall_get_nr(current, regs);
@@ -390,8 +394,9 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
 	if (!sys_data)
 		return;
 
-	tracing_generic_entry_update(&rec.entry, 0, 0);
-	rec.entry.type = sys_data->exit_id;
+	tracing_generic_entry_update(&rec.ent, 0, 0);
+	rec.ent.type = sys_data->exit_id;
+	rec.nr = syscall_nr;
 	rec.ret = syscall_get_return_value(current, regs);
 
 	perf_tpcounter_event(sys_data->exit_id, 0, 1, &rec, sizeof(rec));
-- 
1.6.3

[PATCH 2/5] tracing/syscalls: Add fields format for exit events

[Li Zefan]

Add "format" file for syscall exit events:

 # cat events/syscalls/sys_exit_open/format
 name: sys_exit_open
 ID: 344
 format:
         field:unsigned short common_type;       offset:0;       size:2;
         field:unsigned char common_flags;       offset:2;       size:1;
         field:unsigned char common_preempt_count;       offset:3;       size:1;
         field:int common_pid;   offset:4;       size:4;
         field:int common_tgid;  offset:8;       size:4;

         field:int nr;   offset:12;      size:4;
         field:unsigned long ret;        offset:16;      size:4;

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 include/linux/syscalls.h      |    3 ++-
 include/trace/syscall.h       |    6 ++++--
 kernel/trace/trace_syscalls.c |   18 +++++++++++++++++-
 3 files changed, 23 insertions(+), 4 deletions(-)

diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 87d06c1..8d57f77 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -189,7 +189,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.system                 = "syscalls",			\
 		.event                  = &event_syscall_enter,		\
 		.raw_init		= init_enter_##sname,		\
-		.show_format		= ftrace_format_syscall,	\
+		.show_format		= syscall_enter_format,		\
 		.regfunc		= reg_event_syscall_enter,	\
 		.unregfunc		= unreg_event_syscall_enter,	\
 		.data			= "sys"#sname,			\
@@ -225,6 +225,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.system                 = "syscalls",			\
 		.event                  = &event_syscall_exit,		\
 		.raw_init		= init_exit_##sname,		\
+		.show_format		= syscall_exit_format,		\
 		.regfunc		= reg_event_syscall_exit,	\
 		.unregfunc		= unreg_event_syscall_exit,	\
 		.data			= "sys"#sname,			\
diff --git a/include/trace/syscall.h b/include/trace/syscall.h
index 0cb0362..5ce85d7 100644
--- a/include/trace/syscall.h
+++ b/include/trace/syscall.h
@@ -55,8 +55,10 @@ extern int reg_event_syscall_enter(void *ptr);
 extern void unreg_event_syscall_enter(void *ptr);
 extern int reg_event_syscall_exit(void *ptr);
 extern void unreg_event_syscall_exit(void *ptr);
-extern int
-ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s);
+extern int syscall_enter_format(struct ftrace_event_call *call,
+				struct trace_seq *s);
+extern int syscall_exit_format(struct ftrace_event_call *call,
+				struct trace_seq *s);
 enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags);
 enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags);
 #endif
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index d10daf0..7336b6c 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -97,7 +97,7 @@ extern char *__bad_type_size(void);
 		__bad_type_size() :					\
 		#type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
 
-int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
+int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
 {
 	int i;
 	int nr;
@@ -149,6 +149,22 @@ int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 	return ret;
 }
 
+int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
+{
+	int ret;
+	struct syscall_trace_exit trace;
+
+	ret = trace_seq_printf(s,
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n"
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+			       SYSCALL_FIELD(int, nr),
+			       SYSCALL_FIELD(unsigned long, ret));
+	if (!ret)
+		return 0;
+
+	return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n");
+}
+
 void ftrace_syscall_enter(struct pt_regs *regs, long id)
 {
 	struct syscall_trace_enter *entry;
-- 
1.6.3

[PATCH 3/5] tracing/events: Add ftrace_event_call param to define_fields()

[Li Zefan]

This parameter is needed by syscall events to add define_fields()
handler.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 include/linux/ftrace_event.h |    2 +-
 include/trace/ftrace.h       |    3 +--
 kernel/trace/trace_events.c  |    2 +-
 kernel/trace/trace_export.c  |    5 ++---
 4 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 189806b..35b3a4a 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -118,7 +118,7 @@ struct ftrace_event_call {
 	int			(*raw_init)(void);
 	int			(*show_format)(struct ftrace_event_call *call,
 					       struct trace_seq *s);
-	int			(*define_fields)(void);
+	int			(*define_fields)(struct ftrace_event_call *);
 	struct list_head	fields;
 	int			filter_active;
 	struct event_filter	*filter;
diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h
index b250b06..4e81c9b 100644
--- a/include/trace/ftrace.h
+++ b/include/trace/ftrace.h
@@ -294,10 +294,9 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags)	\
 #undef TRACE_EVENT
 #define TRACE_EVENT(call, proto, args, tstruct, func, print)		\
 int									\
-ftrace_define_fields_##call(void)					\
+ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 {									\
 	struct ftrace_raw_##call field;					\
-	struct ftrace_event_call *event_call = &event_##call;		\
 	int ret;							\
 									\
 	__common_field(int, type, 1);					\
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index b568ade..af8fb8e 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -941,7 +941,7 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events,
 					  id);
 
 	if (call->define_fields) {
-		ret = call->define_fields();
+		ret = call->define_fields(call);
 		if (ret < 0) {
 			pr_warning("Could not initialize trace point"
 				   " events/%s\n", call->name);
diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c
index 956d4bc..cf2c752 100644
--- a/kernel/trace/trace_export.c
+++ b/kernel/trace/trace_export.c
@@ -119,7 +119,7 @@ ftrace_format_##call(struct ftrace_event_call *unused,			\
 
 #undef TRACE_EVENT_FORMAT
 #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt)	\
-int ftrace_define_fields_##call(void);					\
+int ftrace_define_fields_##call(struct ftrace_event_call *event_call);	\
 static int ftrace_raw_init_event_##call(void);				\
 									\
 struct ftrace_event_call __used						\
@@ -184,9 +184,8 @@ __attribute__((section("_ftrace_events"))) event_##call = {		\
 #undef TRACE_EVENT_FORMAT
 #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt)	\
 int									\
-ftrace_define_fields_##call(void)					\
+ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 {									\
-	struct ftrace_event_call *event_call = &event_##call;		\
 	struct args field;						\
 	int ret;							\
 									\
-- 
1.6.3

[PATCH 4/5] tracing/events: Add trace_define_common_fields()

[Li Zefan]

Extract duplicate code. Also prepare for the later patch.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 include/linux/ftrace_event.h |    8 +-------
 include/trace/ftrace.h       |    8 +++-----
 kernel/trace/trace_events.c  |   22 ++++++++++++++++++++++
 kernel/trace/trace_export.c  |    8 +++-----
 4 files changed, 29 insertions(+), 17 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 35b3a4a..427cbae 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -142,6 +142,7 @@ extern int filter_current_check_discard(struct ftrace_event_call *call,
 
 extern int trace_define_field(struct ftrace_event_call *call, char *type,
 			      char *name, int offset, int size, int is_signed);
+extern int trace_define_common_fields(struct ftrace_event_call *call);
 
 #define is_signed_type(type)	(((type)(-1)) < 0)
 
@@ -166,11 +167,4 @@ do {									\
 		__trace_printk(ip, fmt, ##args);			\
 } while (0)
 
-#define __common_field(type, item, is_signed)				\
-	ret = trace_define_field(event_call, #type, "common_" #item,	\
-				 offsetof(typeof(field.ent), item),	\
-				 sizeof(field.ent.item), is_signed);	\
-	if (ret)							\
-		return ret;
-
 #endif /* _LINUX_FTRACE_EVENT_H */
diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h
index 4e81c9b..1274002 100644
--- a/include/trace/ftrace.h
+++ b/include/trace/ftrace.h
@@ -299,11 +299,9 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 	struct ftrace_raw_##call field;					\
 	int ret;							\
 									\
-	__common_field(int, type, 1);					\
-	__common_field(unsigned char, flags, 0);			\
-	__common_field(unsigned char, preempt_count, 0);		\
-	__common_field(int, pid, 1);					\
-	__common_field(int, tgid, 1);					\
+	ret = trace_define_common_fields(event_call);			\
+	if (ret)							\
+		return ret;						\
 									\
 	tstruct;							\
 									\
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index af8fb8e..9c7ecfb 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -62,6 +62,28 @@ err:
 }
 EXPORT_SYMBOL_GPL(trace_define_field);
 
+#define __common_field(type, item)					\
+	ret = trace_define_field(call, #type, "common_" #item,		\
+				 offsetof(typeof(ent), item),		\
+				 sizeof(ent.item),			\
+				 is_signed_type(type));			\
+	if (ret)							\
+		return ret;
+
+int trace_define_common_fields(struct ftrace_event_call *call)
+{
+	int ret;
+	struct trace_entry ent;
+
+	__common_field(unsigned short, type);
+	__common_field(unsigned char, flags);
+	__common_field(unsigned char, preempt_count);
+	__common_field(int, pid);
+	__common_field(int, tgid);
+
+	return ret;
+}
+
 #ifdef CONFIG_MODULES
 
 static void trace_destroy_fields(struct ftrace_event_call *call)
diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c
index cf2c752..7087530 100644
--- a/kernel/trace/trace_export.c
+++ b/kernel/trace/trace_export.c
@@ -189,11 +189,9 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 	struct args field;						\
 	int ret;							\
 									\
-	__common_field(unsigned char, type, 0);				\
-	__common_field(unsigned char, flags, 0);			\
-	__common_field(unsigned char, preempt_count, 0);		\
-	__common_field(int, pid, 1);					\
-	__common_field(int, tgid, 1);					\
+	ret = trace_define_common_fields(event_call);			\
+	if (ret)							\
+		return ret;						\
 									\
 	tstruct;							\
 									\
-- 
1.6.3

[PATCH 5/5] tracing/syscalls: Add filtering support

[Li Zefan]

Add filtering support for syscall events:

 # echo 'mode == 0666' > events/syscalls/sys_enter_open
 # echo 'ret == 0' > events/syscalls/sys_exit_open
 # echo 1 > events/syscalls/sys_enter_open
 # echo 1 > events/syscalls/sys_exit_open
 # cat trace
 ...
   modprobe-3084 [001] 117.463140: sys_open(filename: 917d3e8, flags: 0, mode: 1b6) 
   modprobe-3084 [001] 117.463176: sys_open -> 0x0
       less-3086 [001] 117.510455: sys_open(filename: 9c6bdb8, flags: 8000, mode: 1b6) 
   sendmail-2574 [001] 122.145840: sys_open(filename: b807a365, flags: 0, mode: 1b6) 
 ...


Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 include/linux/ftrace_event.h  |    5 ++-
 include/linux/syscalls.h      |   16 +++++++++----
 include/trace/syscall.h       |    7 +++++
 kernel/trace/trace_events.c   |    5 ++-
 kernel/trace/trace_syscalls.c |   51 +++++++++++++++++++++++++++++++++++++---
 5 files changed, 71 insertions(+), 13 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 07f5cd1..a010429 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -140,8 +140,9 @@ extern int filter_current_check_discard(struct ftrace_event_call *call,
 					void *rec,
 					struct ring_buffer_event *event);
 
-extern int trace_define_field(struct ftrace_event_call *call, char *type,
-			      char *name, int offset, int size, int is_signed);
+extern int trace_define_field(struct ftrace_event_call *call,
+			      const char *type, const char *name,
+			      int offset, int size, int is_signed);
 extern int trace_define_common_fields(struct ftrace_event_call *call);
 
 #define is_signed_type(type)	(((type)(-1)) < 0)
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 8d57f77..f124c89 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -190,6 +190,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.event                  = &event_syscall_enter,		\
 		.raw_init		= init_enter_##sname,		\
 		.show_format		= syscall_enter_format,		\
+		.define_fields		= syscall_enter_define_fields,	\
 		.regfunc		= reg_event_syscall_enter,	\
 		.unregfunc		= unreg_event_syscall_enter,	\
 		.data			= "sys"#sname,			\
@@ -226,6 +227,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.event                  = &event_syscall_exit,		\
 		.raw_init		= init_exit_##sname,		\
 		.show_format		= syscall_exit_format,		\
+		.define_fields		= syscall_exit_define_fields,	\
 		.regfunc		= reg_event_syscall_exit,	\
 		.unregfunc		= unreg_event_syscall_exit,	\
 		.data			= "sys"#sname,			\
@@ -233,6 +235,8 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 	}
 
 #define SYSCALL_METADATA(sname, nb)				\
+	SYSCALL_TRACE_ENTER_EVENT(sname);			\
+	SYSCALL_TRACE_EXIT_EVENT(sname);			\
 	static const struct syscall_metadata __used		\
 	  __attribute__((__aligned__(4)))			\
 	  __attribute__((section("__syscalls_metadata")))	\
@@ -241,20 +245,22 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.nb_args 	= nb,				\
 		.types		= types_##sname,		\
 		.args		= args_##sname,			\
-	};							\
-	SYSCALL_TRACE_ENTER_EVENT(sname);			\
-	SYSCALL_TRACE_EXIT_EVENT(sname);
+		.enter_event	= &event_enter_##sname,		\
+		.exit_event	= &event_exit_##sname,		\
+	};
 
 #define SYSCALL_DEFINE0(sname)					\
+	SYSCALL_TRACE_ENTER_EVENT(_##sname);			\
+	SYSCALL_TRACE_EXIT_EVENT(_##sname);			\
 	static const struct syscall_metadata __used		\
 	  __attribute__((__aligned__(4)))			\
 	  __attribute__((section("__syscalls_metadata")))	\
 	  __syscall_meta_##sname = {				\
 		.name 		= "sys_"#sname,			\
 		.nb_args 	= 0,				\
+		.enter_event	= &event_enter__##sname,	\
+		.exit_event	= &event_exit__##sname,		\
 	};							\
-	SYSCALL_TRACE_ENTER_EVENT(_##sname);			\
-	SYSCALL_TRACE_EXIT_EVENT(_##sname);			\
 	asmlinkage long sys_##sname(void)
 #else
 #define SYSCALL_DEFINE0(name)	   asmlinkage long sys_##name(void)
diff --git a/include/trace/syscall.h b/include/trace/syscall.h
index 5ce85d7..9661dd4 100644
--- a/include/trace/syscall.h
+++ b/include/trace/syscall.h
@@ -34,6 +34,8 @@ DECLARE_TRACE_WITH_CALLBACK(syscall_exit,
  * @args: list of args as strings (args[i] matches types[i])
  * @enter_id: associated ftrace enter event id
  * @exit_id: associated ftrace exit event id
+ * @enter_event: associated syscall_enter trace event
+ * @exit_event: associated syscall_exit trace event
  */
 struct syscall_metadata {
 	const char	*name;
@@ -42,6 +44,9 @@ struct syscall_metadata {
 	const char	**args;
 	int		enter_id;
 	int		exit_id;
+
+	struct ftrace_event_call *enter_event;
+	struct ftrace_event_call *exit_event;
 };
 
 #ifdef CONFIG_FTRACE_SYSCALLS
@@ -59,6 +64,8 @@ extern int syscall_enter_format(struct ftrace_event_call *call,
 				struct trace_seq *s);
 extern int syscall_exit_format(struct ftrace_event_call *call,
 				struct trace_seq *s);
+extern int syscall_enter_define_fields(struct ftrace_event_call *call);
+extern int syscall_exit_define_fields(struct ftrace_event_call *call);
 enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags);
 enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags);
 #endif
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 9c7ecfb..79d3520 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -27,8 +27,8 @@ DEFINE_MUTEX(event_mutex);
 
 LIST_HEAD(ftrace_events);
 
-int trace_define_field(struct ftrace_event_call *call, char *type,
-		       char *name, int offset, int size, int is_signed)
+int trace_define_field(struct ftrace_event_call *call, const char *type,
+		       const char *name, int offset, int size, int is_signed)
 {
 	struct ftrace_event_field *field;
 
@@ -83,6 +83,7 @@ int trace_define_common_fields(struct ftrace_event_call *call)
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(trace_define_common_fields);
 
 #ifdef CONFIG_MODULES
 
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index ce44cdd..495c639 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -163,6 +163,49 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
 	return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n");
 }
 
+int syscall_enter_define_fields(struct ftrace_event_call *call)
+{
+	struct syscall_trace_enter trace;
+	struct syscall_metadata *meta;
+	int ret;
+	int nr;
+	int i;
+	int offset = offsetof(typeof(trace), args);
+
+	nr = syscall_name_to_nr(call->data);
+	meta = syscall_nr_to_meta(nr);
+
+	if (!meta)
+		return 0;
+
+	ret = trace_define_common_fields(call);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < meta->nb_args; i++) {
+		ret = trace_define_field(call, meta->types[i],
+					 meta->args[i], offset,
+					 sizeof(unsigned long), 0);
+		offset += sizeof(unsigned long);
+	}
+
+	return ret;
+}
+
+int syscall_exit_define_fields(struct ftrace_event_call *call)
+{
+	struct syscall_trace_exit trace;
+	int ret;
+
+	ret = trace_define_common_fields(call);
+	if (ret)
+		return ret;
+
+	ret = trace_define_field(call, SYSCALL_FIELD(unsigned long, ret), 0);
+
+	return ret;
+}
+
 void ftrace_syscall_enter(struct pt_regs *regs, long id)
 {
 	struct syscall_trace_enter *entry;
@@ -190,8 +233,8 @@ void ftrace_syscall_enter(struct pt_regs *regs, long id)
 	entry->nr = syscall_nr;
 	syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args);
 
-	trace_current_buffer_unlock_commit(event, 0, 0);
-	trace_wake_up();
+	if (!filter_current_check_discard(sys_data->enter_event, entry, event))
+		trace_current_buffer_unlock_commit(event, 0, 0);
 }
 
 void ftrace_syscall_exit(struct pt_regs *regs, long ret)
@@ -218,8 +261,8 @@ void ftrace_syscall_exit(struct pt_regs *regs, long ret)
 	entry->nr = syscall_nr;
 	entry->ret = syscall_get_return_value(current, regs);
 
-	trace_current_buffer_unlock_commit(event, 0, 0);
-	trace_wake_up();
+	if (!filter_current_check_discard(sys_data->exit_event, entry, event))
+		trace_current_buffer_unlock_commit(event, 0, 0);
 }
 
 int reg_event_syscall_enter(void *ptr)
-- 
1.6.3

[tip:tracing/core] tracing/syscalls: Fix fields format for enter events

[tip-bot for Li Zefan]

Commit-ID:  e6971969c331caa5c3c88cbd1be4f465b3355452
Gitweb:     http://git.kernel.org/tip/e6971969c331caa5c3c88cbd1be4f465b3355452
Author:     Li Zefan <lizf@cn.fujitsu.com>
AuthorDate: Wed, 19 Aug 2009 15:52:25 +0800
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 19 Aug 2009 15:02:22 +0200

tracing/syscalls: Fix fields format for enter events

The "format" file of a trace event is originally for parsers to
parse ftrace binary output.

But the "format" file of a syscall event can only be used by
perfcounter, because it describes the format of struct
syscall_enter_record not struct syscall_trace_enter.

To fix this, we remove struct syscall_enter_record, and then
struct syscall_trace_enter will be used by both perf profile
and ftrace.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4A8BAF39.1030404@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 kernel/trace/trace_syscalls.c |   51 ++++++++++++++++++++++------------------
 1 files changed, 28 insertions(+), 23 deletions(-)

diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index f130dac..d10daf0 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -90,26 +90,39 @@ print_syscall_exit(struct trace_iterator *iter, int flags)
 	return TRACE_TYPE_HANDLED;
 }
 
+extern char *__bad_type_size(void);
+
+#define SYSCALL_FIELD(type, name)					\
+	sizeof(type) != sizeof(trace.name) ?				\
+		__bad_type_size() :					\
+		#type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
+
 int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 {
 	int i;
 	int nr;
-	int ret = 0;
+	int ret;
 	struct syscall_metadata *entry;
-	int offset = sizeof(struct trace_entry);
+	struct syscall_trace_enter trace;
+	int offset = offsetof(struct syscall_trace_enter, args);
 
-	nr = syscall_name_to_nr((char *)call->data);
+	nr = syscall_name_to_nr(call->data);
 	entry = syscall_nr_to_meta(nr);
 
 	if (!entry)
-		return ret;
+		return 0;
+
+	ret = trace_seq_printf(s, "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+			       SYSCALL_FIELD(int, nr));
+	if (!ret)
+		return 0;
 
 	for (i = 0; i < entry->nb_args; i++) {
 		ret = trace_seq_printf(s, "\tfield:%s %s;", entry->types[i],
 				        entry->args[i]);
 		if (!ret)
 			return 0;
-		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%lu;\n", offset,
+		ret = trace_seq_printf(s, "\toffset:%d;\tsize:%zu;\n", offset,
 				       sizeof(unsigned long));
 		if (!ret)
 			return 0;
@@ -118,7 +131,7 @@ int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 
 	trace_seq_printf(s, "\nprint fmt: \"");
 	for (i = 0; i < entry->nb_args; i++) {
-		ret = trace_seq_printf(s, "%s: 0x%%0%lulx%s", entry->args[i],
+		ret = trace_seq_printf(s, "%s: 0x%%0%zulx%s", entry->args[i],
 				        sizeof(unsigned long),
 					i == entry->nb_args - 1 ? "\", " : ", ");
 		if (!ret)
@@ -287,16 +300,6 @@ struct trace_event event_syscall_exit = {
 
 #ifdef CONFIG_EVENT_PROFILE
 
-struct syscall_enter_record {
-	struct trace_entry	entry;
-	unsigned long		args[0];
-};
-
-struct syscall_exit_record {
-	struct trace_entry	entry;
-	unsigned long		ret;
-};
-
 static DECLARE_BITMAP(enabled_prof_enter_syscalls, FTRACE_SYSCALL_MAX);
 static DECLARE_BITMAP(enabled_prof_exit_syscalls, FTRACE_SYSCALL_MAX);
 static int sys_prof_refcount_enter;
@@ -304,7 +307,7 @@ static int sys_prof_refcount_exit;
 
 static void prof_syscall_enter(struct pt_regs *regs, long id)
 {
-	struct syscall_enter_record *rec;
+	struct syscall_trace_enter *rec;
 	struct syscall_metadata *sys_data;
 	int syscall_nr;
 	int size;
@@ -328,9 +331,10 @@ static void prof_syscall_enter(struct pt_regs *regs, long id)
 		/* zero the dead bytes from align to not leak stack to user */
 		*(u64 *)(&raw_data[size - sizeof(u64)]) = 0ULL;
 
-		rec = (struct syscall_enter_record *) raw_data;
-		tracing_generic_entry_update(&rec->entry, 0, 0);
-		rec->entry.type = sys_data->enter_id;
+		rec = (struct syscall_trace_enter *) raw_data;
+		tracing_generic_entry_update(&rec->ent, 0, 0);
+		rec->ent.type = sys_data->enter_id;
+		rec->nr = syscall_nr;
 		syscall_get_arguments(current, regs, 0, sys_data->nb_args,
 				       (unsigned long *)&rec->args);
 		perf_tpcounter_event(sys_data->enter_id, 0, 1, rec, size);
@@ -379,7 +383,7 @@ void unreg_prof_syscall_enter(char *name)
 static void prof_syscall_exit(struct pt_regs *regs, long ret)
 {
 	struct syscall_metadata *sys_data;
-	struct syscall_exit_record rec;
+	struct syscall_trace_exit rec;
 	int syscall_nr;
 
 	syscall_nr = syscall_get_nr(current, regs);
@@ -390,8 +394,9 @@ static void prof_syscall_exit(struct pt_regs *regs, long ret)
 	if (!sys_data)
 		return;
 
-	tracing_generic_entry_update(&rec.entry, 0, 0);
-	rec.entry.type = sys_data->exit_id;
+	tracing_generic_entry_update(&rec.ent, 0, 0);
+	rec.ent.type = sys_data->exit_id;
+	rec.nr = syscall_nr;
 	rec.ret = syscall_get_return_value(current, regs);
 
 	perf_tpcounter_event(sys_data->exit_id, 0, 1, &rec, sizeof(rec));

[tip:tracing/core] tracing/syscalls: Add fields format for exit events

[tip-bot for Li Zefan]

Commit-ID:  10a5b66f625904ad5a2867cf7a28073e1236ff32
Gitweb:     http://git.kernel.org/tip/10a5b66f625904ad5a2867cf7a28073e1236ff32
Author:     Li Zefan <lizf@cn.fujitsu.com>
AuthorDate: Wed, 19 Aug 2009 15:53:05 +0800
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 19 Aug 2009 15:02:23 +0200

tracing/syscalls: Add fields format for exit events

Add "format" file for syscall exit events:

 # cat events/syscalls/sys_exit_open/format
 name: sys_exit_open
 ID: 344
 format:
         field:unsigned short common_type;       offset:0;       size:2;
         field:unsigned char common_flags;       offset:2;       size:1;
         field:unsigned char common_preempt_count;       offset:3;       size:1;
         field:int common_pid;   offset:4;       size:4;
         field:int common_tgid;  offset:8;       size:4;

         field:int nr;   offset:12;      size:4;
         field:unsigned long ret;        offset:16;      size:4;

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4A8BAF61.3060307@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 include/linux/syscalls.h      |    3 ++-
 include/trace/syscall.h       |    6 ++++--
 kernel/trace/trace_syscalls.c |   18 +++++++++++++++++-
 3 files changed, 23 insertions(+), 4 deletions(-)

diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 87d06c1..8d57f77 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -189,7 +189,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.system                 = "syscalls",			\
 		.event                  = &event_syscall_enter,		\
 		.raw_init		= init_enter_##sname,		\
-		.show_format		= ftrace_format_syscall,	\
+		.show_format		= syscall_enter_format,		\
 		.regfunc		= reg_event_syscall_enter,	\
 		.unregfunc		= unreg_event_syscall_enter,	\
 		.data			= "sys"#sname,			\
@@ -225,6 +225,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.system                 = "syscalls",			\
 		.event                  = &event_syscall_exit,		\
 		.raw_init		= init_exit_##sname,		\
+		.show_format		= syscall_exit_format,		\
 		.regfunc		= reg_event_syscall_exit,	\
 		.unregfunc		= unreg_event_syscall_exit,	\
 		.data			= "sys"#sname,			\
diff --git a/include/trace/syscall.h b/include/trace/syscall.h
index 0cb0362..5ce85d7 100644
--- a/include/trace/syscall.h
+++ b/include/trace/syscall.h
@@ -55,8 +55,10 @@ extern int reg_event_syscall_enter(void *ptr);
 extern void unreg_event_syscall_enter(void *ptr);
 extern int reg_event_syscall_exit(void *ptr);
 extern void unreg_event_syscall_exit(void *ptr);
-extern int
-ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s);
+extern int syscall_enter_format(struct ftrace_event_call *call,
+				struct trace_seq *s);
+extern int syscall_exit_format(struct ftrace_event_call *call,
+				struct trace_seq *s);
 enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags);
 enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags);
 #endif
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index d10daf0..7336b6c 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -97,7 +97,7 @@ extern char *__bad_type_size(void);
 		__bad_type_size() :					\
 		#type, #name, offsetof(typeof(trace), name), sizeof(trace.name)
 
-int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
+int syscall_enter_format(struct ftrace_event_call *call, struct trace_seq *s)
 {
 	int i;
 	int nr;
@@ -149,6 +149,22 @@ int ftrace_format_syscall(struct ftrace_event_call *call, struct trace_seq *s)
 	return ret;
 }
 
+int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
+{
+	int ret;
+	struct syscall_trace_exit trace;
+
+	ret = trace_seq_printf(s,
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n"
+			       "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n",
+			       SYSCALL_FIELD(int, nr),
+			       SYSCALL_FIELD(unsigned long, ret));
+	if (!ret)
+		return 0;
+
+	return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n");
+}
+
 void ftrace_syscall_enter(struct pt_regs *regs, long id)
 {
 	struct syscall_trace_enter *entry;

[tip:tracing/core] tracing/events: Add ftrace_event_call param to define_fields()

[tip-bot for Li Zefan]

Commit-ID:  14be96c9716cb8c46dca94bd890defd7856e0734
Gitweb:     http://git.kernel.org/tip/14be96c9716cb8c46dca94bd890defd7856e0734
Author:     Li Zefan <lizf@cn.fujitsu.com>
AuthorDate: Wed, 19 Aug 2009 15:53:52 +0800
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 19 Aug 2009 15:02:23 +0200

tracing/events: Add ftrace_event_call param to define_fields()

This parameter is needed by syscall events to add define_fields()
handler.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4A8BAF90.6060801@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 include/linux/ftrace_event.h |    2 +-
 include/trace/ftrace.h       |    3 +--
 kernel/trace/trace_events.c  |    2 +-
 kernel/trace/trace_export.c  |    5 ++---
 4 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 189806b..35b3a4a 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -118,7 +118,7 @@ struct ftrace_event_call {
 	int			(*raw_init)(void);
 	int			(*show_format)(struct ftrace_event_call *call,
 					       struct trace_seq *s);
-	int			(*define_fields)(void);
+	int			(*define_fields)(struct ftrace_event_call *);
 	struct list_head	fields;
 	int			filter_active;
 	struct event_filter	*filter;
diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h
index b250b06..4e81c9b 100644
--- a/include/trace/ftrace.h
+++ b/include/trace/ftrace.h
@@ -294,10 +294,9 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags)	\
 #undef TRACE_EVENT
 #define TRACE_EVENT(call, proto, args, tstruct, func, print)		\
 int									\
-ftrace_define_fields_##call(void)					\
+ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 {									\
 	struct ftrace_raw_##call field;					\
-	struct ftrace_event_call *event_call = &event_##call;		\
 	int ret;							\
 									\
 	__common_field(int, type, 1);					\
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index b568ade..af8fb8e 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -941,7 +941,7 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events,
 					  id);
 
 	if (call->define_fields) {
-		ret = call->define_fields();
+		ret = call->define_fields(call);
 		if (ret < 0) {
 			pr_warning("Could not initialize trace point"
 				   " events/%s\n", call->name);
diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c
index 956d4bc..cf2c752 100644
--- a/kernel/trace/trace_export.c
+++ b/kernel/trace/trace_export.c
@@ -119,7 +119,7 @@ ftrace_format_##call(struct ftrace_event_call *unused,			\
 
 #undef TRACE_EVENT_FORMAT
 #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt)	\
-int ftrace_define_fields_##call(void);					\
+int ftrace_define_fields_##call(struct ftrace_event_call *event_call);	\
 static int ftrace_raw_init_event_##call(void);				\
 									\
 struct ftrace_event_call __used						\
@@ -184,9 +184,8 @@ __attribute__((section("_ftrace_events"))) event_##call = {		\
 #undef TRACE_EVENT_FORMAT
 #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt)	\
 int									\
-ftrace_define_fields_##call(void)					\
+ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 {									\
-	struct ftrace_event_call *event_call = &event_##call;		\
 	struct args field;						\
 	int ret;							\
 									\

[tip:tracing/core] tracing/events: Add trace_define_common_fields()

[tip-bot for Li Zefan]

Commit-ID:  e647d6b314266adb904d4b84973eda0afa856946
Gitweb:     http://git.kernel.org/tip/e647d6b314266adb904d4b84973eda0afa856946
Author:     Li Zefan <lizf@cn.fujitsu.com>
AuthorDate: Wed, 19 Aug 2009 15:54:32 +0800
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 19 Aug 2009 15:02:24 +0200

tracing/events: Add trace_define_common_fields()

Extract duplicate code. Also prepare for the later patch.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4A8BAFB8.1010304@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 include/linux/ftrace_event.h |    8 +-------
 include/trace/ftrace.h       |    8 +++-----
 kernel/trace/trace_events.c  |   22 ++++++++++++++++++++++
 kernel/trace/trace_export.c  |    8 +++-----
 4 files changed, 29 insertions(+), 17 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 35b3a4a..427cbae 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -142,6 +142,7 @@ extern int filter_current_check_discard(struct ftrace_event_call *call,
 
 extern int trace_define_field(struct ftrace_event_call *call, char *type,
 			      char *name, int offset, int size, int is_signed);
+extern int trace_define_common_fields(struct ftrace_event_call *call);
 
 #define is_signed_type(type)	(((type)(-1)) < 0)
 
@@ -166,11 +167,4 @@ do {									\
 		__trace_printk(ip, fmt, ##args);			\
 } while (0)
 
-#define __common_field(type, item, is_signed)				\
-	ret = trace_define_field(event_call, #type, "common_" #item,	\
-				 offsetof(typeof(field.ent), item),	\
-				 sizeof(field.ent.item), is_signed);	\
-	if (ret)							\
-		return ret;
-
 #endif /* _LINUX_FTRACE_EVENT_H */
diff --git a/include/trace/ftrace.h b/include/trace/ftrace.h
index 4e81c9b..1274002 100644
--- a/include/trace/ftrace.h
+++ b/include/trace/ftrace.h
@@ -299,11 +299,9 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 	struct ftrace_raw_##call field;					\
 	int ret;							\
 									\
-	__common_field(int, type, 1);					\
-	__common_field(unsigned char, flags, 0);			\
-	__common_field(unsigned char, preempt_count, 0);		\
-	__common_field(int, pid, 1);					\
-	__common_field(int, tgid, 1);					\
+	ret = trace_define_common_fields(event_call);			\
+	if (ret)							\
+		return ret;						\
 									\
 	tstruct;							\
 									\
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index af8fb8e..9c7ecfb 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -62,6 +62,28 @@ err:
 }
 EXPORT_SYMBOL_GPL(trace_define_field);
 
+#define __common_field(type, item)					\
+	ret = trace_define_field(call, #type, "common_" #item,		\
+				 offsetof(typeof(ent), item),		\
+				 sizeof(ent.item),			\
+				 is_signed_type(type));			\
+	if (ret)							\
+		return ret;
+
+int trace_define_common_fields(struct ftrace_event_call *call)
+{
+	int ret;
+	struct trace_entry ent;
+
+	__common_field(unsigned short, type);
+	__common_field(unsigned char, flags);
+	__common_field(unsigned char, preempt_count);
+	__common_field(int, pid);
+	__common_field(int, tgid);
+
+	return ret;
+}
+
 #ifdef CONFIG_MODULES
 
 static void trace_destroy_fields(struct ftrace_event_call *call)
diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c
index cf2c752..7087530 100644
--- a/kernel/trace/trace_export.c
+++ b/kernel/trace/trace_export.c
@@ -189,11 +189,9 @@ ftrace_define_fields_##call(struct ftrace_event_call *event_call)	\
 	struct args field;						\
 	int ret;							\
 									\
-	__common_field(unsigned char, type, 0);				\
-	__common_field(unsigned char, flags, 0);			\
-	__common_field(unsigned char, preempt_count, 0);		\
-	__common_field(int, pid, 1);					\
-	__common_field(int, tgid, 1);					\
+	ret = trace_define_common_fields(event_call);			\
+	if (ret)							\
+		return ret;						\
 									\
 	tstruct;							\
 									\

[tip:tracing/core] tracing/syscalls: Add filtering support

[tip-bot for Li Zefan]

Commit-ID:  540b7b8d65575c80162f2a0f38e1d313c92a6042
Gitweb:     http://git.kernel.org/tip/540b7b8d65575c80162f2a0f38e1d313c92a6042
Author:     Li Zefan <lizf@cn.fujitsu.com>
AuthorDate: Wed, 19 Aug 2009 15:54:51 +0800
Committer:  Ingo Molnar <mingo@elte.hu>
CommitDate: Wed, 19 Aug 2009 15:02:24 +0200

tracing/syscalls: Add filtering support

Add filtering support for syscall events:

 # echo 'mode == 0666' > events/syscalls/sys_enter_open
 # echo 'ret == 0' > events/syscalls/sys_exit_open
 # echo 1 > events/syscalls/sys_enter_open
 # echo 1 > events/syscalls/sys_exit_open
 # cat trace
 ...
   modprobe-3084 [001] 117.463140: sys_open(filename: 917d3e8, flags: 0, mode: 1b6)
   modprobe-3084 [001] 117.463176: sys_open -> 0x0
       less-3086 [001] 117.510455: sys_open(filename: 9c6bdb8, flags: 8000, mode: 1b6)
   sendmail-2574 [001] 122.145840: sys_open(filename: b807a365, flags: 0, mode: 1b6)
 ...

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Jason Baron <jbaron@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <4A8BAFCB.1040006@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>


---
 include/linux/ftrace_event.h  |    5 ++-
 include/linux/syscalls.h      |   16 +++++++++----
 include/trace/syscall.h       |    7 +++++
 kernel/trace/trace_events.c   |    5 ++-
 kernel/trace/trace_syscalls.c |   51 +++++++++++++++++++++++++++++++++++++---
 5 files changed, 71 insertions(+), 13 deletions(-)

diff --git a/include/linux/ftrace_event.h b/include/linux/ftrace_event.h
index 427cbae..df5b085 100644
--- a/include/linux/ftrace_event.h
+++ b/include/linux/ftrace_event.h
@@ -140,8 +140,9 @@ extern int filter_current_check_discard(struct ftrace_event_call *call,
 					void *rec,
 					struct ring_buffer_event *event);
 
-extern int trace_define_field(struct ftrace_event_call *call, char *type,
-			      char *name, int offset, int size, int is_signed);
+extern int trace_define_field(struct ftrace_event_call *call,
+			      const char *type, const char *name,
+			      int offset, int size, int is_signed);
 extern int trace_define_common_fields(struct ftrace_event_call *call);
 
 #define is_signed_type(type)	(((type)(-1)) < 0)
diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h
index 8d57f77..f124c89 100644
--- a/include/linux/syscalls.h
+++ b/include/linux/syscalls.h
@@ -190,6 +190,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.event                  = &event_syscall_enter,		\
 		.raw_init		= init_enter_##sname,		\
 		.show_format		= syscall_enter_format,		\
+		.define_fields		= syscall_enter_define_fields,	\
 		.regfunc		= reg_event_syscall_enter,	\
 		.unregfunc		= unreg_event_syscall_enter,	\
 		.data			= "sys"#sname,			\
@@ -226,6 +227,7 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.event                  = &event_syscall_exit,		\
 		.raw_init		= init_exit_##sname,		\
 		.show_format		= syscall_exit_format,		\
+		.define_fields		= syscall_exit_define_fields,	\
 		.regfunc		= reg_event_syscall_exit,	\
 		.unregfunc		= unreg_event_syscall_exit,	\
 		.data			= "sys"#sname,			\
@@ -233,6 +235,8 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 	}
 
 #define SYSCALL_METADATA(sname, nb)				\
+	SYSCALL_TRACE_ENTER_EVENT(sname);			\
+	SYSCALL_TRACE_EXIT_EVENT(sname);			\
 	static const struct syscall_metadata __used		\
 	  __attribute__((__aligned__(4)))			\
 	  __attribute__((section("__syscalls_metadata")))	\
@@ -241,20 +245,22 @@ static void prof_sysexit_disable_##sname(struct ftrace_event_call *event_call) \
 		.nb_args 	= nb,				\
 		.types		= types_##sname,		\
 		.args		= args_##sname,			\
-	};							\
-	SYSCALL_TRACE_ENTER_EVENT(sname);			\
-	SYSCALL_TRACE_EXIT_EVENT(sname);
+		.enter_event	= &event_enter_##sname,		\
+		.exit_event	= &event_exit_##sname,		\
+	};
 
 #define SYSCALL_DEFINE0(sname)					\
+	SYSCALL_TRACE_ENTER_EVENT(_##sname);			\
+	SYSCALL_TRACE_EXIT_EVENT(_##sname);			\
 	static const struct syscall_metadata __used		\
 	  __attribute__((__aligned__(4)))			\
 	  __attribute__((section("__syscalls_metadata")))	\
 	  __syscall_meta_##sname = {				\
 		.name 		= "sys_"#sname,			\
 		.nb_args 	= 0,				\
+		.enter_event	= &event_enter__##sname,	\
+		.exit_event	= &event_exit__##sname,		\
 	};							\
-	SYSCALL_TRACE_ENTER_EVENT(_##sname);			\
-	SYSCALL_TRACE_EXIT_EVENT(_##sname);			\
 	asmlinkage long sys_##sname(void)
 #else
 #define SYSCALL_DEFINE0(name)	   asmlinkage long sys_##name(void)
diff --git a/include/trace/syscall.h b/include/trace/syscall.h
index 5ce85d7..9661dd4 100644
--- a/include/trace/syscall.h
+++ b/include/trace/syscall.h
@@ -34,6 +34,8 @@ DECLARE_TRACE_WITH_CALLBACK(syscall_exit,
  * @args: list of args as strings (args[i] matches types[i])
  * @enter_id: associated ftrace enter event id
  * @exit_id: associated ftrace exit event id
+ * @enter_event: associated syscall_enter trace event
+ * @exit_event: associated syscall_exit trace event
  */
 struct syscall_metadata {
 	const char	*name;
@@ -42,6 +44,9 @@ struct syscall_metadata {
 	const char	**args;
 	int		enter_id;
 	int		exit_id;
+
+	struct ftrace_event_call *enter_event;
+	struct ftrace_event_call *exit_event;
 };
 
 #ifdef CONFIG_FTRACE_SYSCALLS
@@ -59,6 +64,8 @@ extern int syscall_enter_format(struct ftrace_event_call *call,
 				struct trace_seq *s);
 extern int syscall_exit_format(struct ftrace_event_call *call,
 				struct trace_seq *s);
+extern int syscall_enter_define_fields(struct ftrace_event_call *call);
+extern int syscall_exit_define_fields(struct ftrace_event_call *call);
 enum print_line_t print_syscall_enter(struct trace_iterator *iter, int flags);
 enum print_line_t print_syscall_exit(struct trace_iterator *iter, int flags);
 #endif
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 9c7ecfb..79d3520 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -27,8 +27,8 @@ DEFINE_MUTEX(event_mutex);
 
 LIST_HEAD(ftrace_events);
 
-int trace_define_field(struct ftrace_event_call *call, char *type,
-		       char *name, int offset, int size, int is_signed)
+int trace_define_field(struct ftrace_event_call *call, const char *type,
+		       const char *name, int offset, int size, int is_signed)
 {
 	struct ftrace_event_field *field;
 
@@ -83,6 +83,7 @@ int trace_define_common_fields(struct ftrace_event_call *call)
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(trace_define_common_fields);
 
 #ifdef CONFIG_MODULES
 
diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c
index 7336b6c..28e4dae 100644
--- a/kernel/trace/trace_syscalls.c
+++ b/kernel/trace/trace_syscalls.c
@@ -165,6 +165,49 @@ int syscall_exit_format(struct ftrace_event_call *call, struct trace_seq *s)
 	return trace_seq_printf(s, "\nprint fmt: \"0x%%lx\", REC->ret\n");
 }
 
+int syscall_enter_define_fields(struct ftrace_event_call *call)
+{
+	struct syscall_trace_enter trace;
+	struct syscall_metadata *meta;
+	int ret;
+	int nr;
+	int i;
+	int offset = offsetof(typeof(trace), args);
+
+	nr = syscall_name_to_nr(call->data);
+	meta = syscall_nr_to_meta(nr);
+
+	if (!meta)
+		return 0;
+
+	ret = trace_define_common_fields(call);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < meta->nb_args; i++) {
+		ret = trace_define_field(call, meta->types[i],
+					 meta->args[i], offset,
+					 sizeof(unsigned long), 0);
+		offset += sizeof(unsigned long);
+	}
+
+	return ret;
+}
+
+int syscall_exit_define_fields(struct ftrace_event_call *call)
+{
+	struct syscall_trace_exit trace;
+	int ret;
+
+	ret = trace_define_common_fields(call);
+	if (ret)
+		return ret;
+
+	ret = trace_define_field(call, SYSCALL_FIELD(unsigned long, ret), 0);
+
+	return ret;
+}
+
 void ftrace_syscall_enter(struct pt_regs *regs, long id)
 {
 	struct syscall_trace_enter *entry;
@@ -192,8 +235,8 @@ void ftrace_syscall_enter(struct pt_regs *regs, long id)
 	entry->nr = syscall_nr;
 	syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args);
 
-	trace_current_buffer_unlock_commit(event, 0, 0);
-	trace_wake_up();
+	if (!filter_current_check_discard(sys_data->enter_event, entry, event))
+		trace_current_buffer_unlock_commit(event, 0, 0);
 }
 
 void ftrace_syscall_exit(struct pt_regs *regs, long ret)
@@ -220,8 +263,8 @@ void ftrace_syscall_exit(struct pt_regs *regs, long ret)
 	entry->nr = syscall_nr;
 	entry->ret = syscall_get_return_value(current, regs);
 
-	trace_current_buffer_unlock_commit(event, 0, 0);
-	trace_wake_up();
+	if (!filter_current_check_discard(sys_data->exit_event, entry, event))
+		trace_current_buffer_unlock_commit(event, 0, 0);
 }
 
 int reg_event_syscall_enter(void *ptr)

Re: [PATCH 0/5] tracing/syscalls: Add filtering support

[Frederic Weisbecker]

On Wed, Aug 19, 2009 at 03:51:50PM +0800, Li Zefan wrote:
> This patchset adds filtering support for syscall events
> 
> For syscall enter events, we can do filtering on syscall
> parameters, and for exit events, filter on syscall return
> value.
> 
> For example, we set 'mode == 0666' (0666 == 0x1b6) to
> sys_enter_open, and set 'ret == 0' to sys_exit_open,
> and here's the output:
> 
>  # echo 'mode == 0666' > events/syscalls/sys_enter_open
>  # echo 'ret == 0' > events/syscalls/sys_exit_open
>  # echo 1 > events/syscalls/sys_enter_open
>  # echo 1 > events/syscalls/sys_exit_open
>  # cat trace
>  ...
>    modprobe-3084 [001] 117.463140: sys_open(filename: 917d3e8, flags: 0, mode: 1b6) 
>    modprobe-3084 [001] 117.463176: sys_open -> 0x0
>        less-3086 [001] 117.510455: sys_open(filename: 9c6bdb8, flags: 8000, mode: 1b6) 
>    sendmail-2574 [001] 122.145840: sys_open(filename: b807a365, flags: 0, mode: 1b6) 
>  ...



Nice! Thanks!


> 
> [PATCH 1/5] tracing/syscalls: Fix fields format for enter events
> [PATCH 2/5] tracing/syscalls: Add fields format for exit events
> [PATCH 3/5] tracing/events: Add ftrace_event_call param to define_fields()
> [PATCH 4/5] tracing/events: Add trace_define_common_fields()
> [PATCH 5/5] tracing/syscalls: Add filtering support
> ---
>  include/linux/ftrace_event.h  |   15 ++----
>  include/linux/syscalls.h      |   19 +++++--
>  include/trace/ftrace.h        |   11 ++---
>  include/trace/syscall.h       |   13 ++++-
>  kernel/trace/trace_events.c   |   29 +++++++++-
>  kernel/trace/trace_export.c   |   13 ++---
>  kernel/trace/trace_syscalls.c |  120 +++++++++++++++++++++++++++++++----------
>  7 files changed, 156 insertions(+), 64 deletions(-)