[net-next-2.6.git] / kernel / trace / trace_syscalls.c

#include <trace/syscall.h>
#include <trace/events/syscalls.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/ftrace.h>
#include <linux/perf_event.h>
#include <asm/syscall.h>

#include "trace_output.h"
#include "trace.h"

static DEFINE_MUTEX(syscall_trace_lock);
static int sys_refcount_enter;
static int sys_refcount_exit;
static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);

extern unsigned long __start_syscalls_metadata[];
extern unsigned long __stop_syscalls_metadata[];

static struct syscall_metadata **syscalls_metadata;

static struct syscall_metadata *find_syscall_meta(unsigned long syscall)
{
	struct syscall_metadata *start;
	struct syscall_metadata *stop;
	char str[KSYM_SYMBOL_LEN];


	start = (struct syscall_metadata *)__start_syscalls_metadata;
	stop = (struct syscall_metadata *)__stop_syscalls_metadata;
	kallsyms_lookup(syscall, NULL, NULL, NULL, str);

	for ( ; start < stop; start++) {
		/*
		 * Only compare after the "sys" prefix. Archs that use
		 * syscall wrappers may have syscalls symbols aliases prefixed
		 * with "SyS" instead of "sys", leading to an unwanted
		 * mismatch.
		 */
		if (start->name && !strcmp(start->name + 3, str + 3))
			return start;
	}
	return NULL;
}

static struct syscall_metadata *syscall_nr_to_meta(int nr)
{
	if (!syscalls_metadata || nr >= NR_syscalls || nr < 0)
		return NULL;

	return syscalls_metadata[nr];
}

enum print_line_t
print_syscall_enter(struct trace_iterator *iter, int flags)
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *ent = iter->ent;
	struct syscall_trace_enter *trace;
	struct syscall_metadata *entry;
	int i, ret, syscall;

	trace = (typeof(trace))ent;
	syscall = trace->nr;
	entry = syscall_nr_to_meta(syscall);

	if (!entry)
		goto end;

	if (entry->enter_event->id != ent->type) {
		WARN_ON_ONCE(1);
		goto end;
	}

	ret = trace_seq_printf(s, "%s(", entry->name);
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	for (i = 0; i < entry->nb_args; i++) {
		/* parameter types */
		if (trace_flags & TRACE_ITER_VERBOSE) {
			ret = trace_seq_printf(s, "%s ", entry->types[i]);
			if (!ret)
				return TRACE_TYPE_PARTIAL_LINE;
		}
		/* parameter values */
		ret = trace_seq_printf(s, "%s: %lx%s", entry->args[i],
				       trace->args[i],
				       i == entry->nb_args - 1 ? "" : ", ");
		if (!ret)
			return TRACE_TYPE_PARTIAL_LINE;
	}

	ret = trace_seq_putc(s, ')');
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

end:
	ret =  trace_seq_putc(s, '\n');
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	return TRACE_TYPE_HANDLED;
}

enum print_line_t
print_syscall_exit(struct trace_iterator *iter, int flags)
{
	struct trace_seq *s = &iter->seq;
	struct trace_entry *ent = iter->ent;
	struct syscall_trace_exit *trace;
	int syscall;
	struct syscall_metadata *entry;
	int ret;

	trace = (typeof(trace))ent;
	syscall = trace->nr;
	entry = syscall_nr_to_meta(syscall);

	if (!entry) {
		trace_seq_printf(s, "\n");
		return TRACE_TYPE_HANDLED;
	}

	if (entry->exit_event->id != ent->type) {
		WARN_ON_ONCE(1);
		return TRACE_TYPE_UNHANDLED;
	}

	ret = trace_seq_printf(s, "%s -> 0x%lx\n", entry->name,
				trace->ret);
	if (!ret)
		return TRACE_TYPE_PARTIAL_LINE;

	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), is_signed_type(type)

static
int  __set_enter_print_fmt(struct syscall_metadata *entry, char *buf, int len)
{
	int i;
	int pos = 0;

	/* When len=0, we just calculate the needed length */
#define LEN_OR_ZERO (len ? len - pos : 0)

	pos += snprintf(buf + pos, LEN_OR_ZERO, "\"");
	for (i = 0; i < entry->nb_args; i++) {
		pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s",
				entry->args[i], sizeof(unsigned long),
				i == entry->nb_args - 1 ? "" : ", ");
	}
	pos += snprintf(buf + pos, LEN_OR_ZERO, "\"");

	for (i = 0; i < entry->nb_args; i++) {
		pos += snprintf(buf + pos, LEN_OR_ZERO,
				", ((unsigned long)(REC->%s))", entry->args[i]);
	}

#undef LEN_OR_ZERO

	/* return the length of print_fmt */
	return pos;
}

static int set_syscall_print_fmt(struct ftrace_event_call *call)
{
	char *print_fmt;
	int len;
	struct syscall_metadata *entry = call->data;

	if (entry->enter_event != call) {
		call->print_fmt = "\"0x%lx\", REC->ret";
		return 0;
	}

	/* First: called with 0 length to calculate the needed length */
	len = __set_enter_print_fmt(entry, NULL, 0);

	print_fmt = kmalloc(len + 1, GFP_KERNEL);
	if (!print_fmt)
		return -ENOMEM;

	/* Second: actually write the @print_fmt */
	__set_enter_print_fmt(entry, print_fmt, len + 1);
	call->print_fmt = print_fmt;

	return 0;
}

static void free_syscall_print_fmt(struct ftrace_event_call *call)
{
	struct syscall_metadata *entry = call->data;

	if (entry->enter_event == call)
		kfree(call->print_fmt);
}

int syscall_enter_define_fields(struct ftrace_event_call *call)
{
	struct syscall_trace_enter trace;
	struct syscall_metadata *meta = call->data;
	int ret;
	int i;
	int offset = offsetof(typeof(trace), args);

	ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
	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,
					 FILTER_OTHER);
		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_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
	if (ret)
		return ret;

	ret = trace_define_field(call, SYSCALL_FIELD(long, ret),
				 FILTER_OTHER);

	return ret;
}

void ftrace_syscall_enter(struct pt_regs *regs, long id)
{
	struct syscall_trace_enter *entry;
	struct syscall_metadata *sys_data;
	struct ring_buffer_event *event;
	struct ring_buffer *buffer;
	int size;
	int syscall_nr;

	syscall_nr = syscall_get_nr(current, regs);
	if (syscall_nr < 0)
		return;
	if (!test_bit(syscall_nr, enabled_enter_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

	size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args;

	event = trace_current_buffer_lock_reserve(&buffer,
			sys_data->enter_event->id, size, 0, 0);
	if (!event)
		return;

	entry = ring_buffer_event_data(event);
	entry->nr = syscall_nr;
	syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args);

	if (!filter_current_check_discard(buffer, sys_data->enter_event,
					  entry, event))
		trace_current_buffer_unlock_commit(buffer, event, 0, 0);
}

void ftrace_syscall_exit(struct pt_regs *regs, long ret)
{
	struct syscall_trace_exit *entry;
	struct syscall_metadata *sys_data;
	struct ring_buffer_event *event;
	struct ring_buffer *buffer;
	int syscall_nr;

	syscall_nr = syscall_get_nr(current, regs);
	if (syscall_nr < 0)
		return;
	if (!test_bit(syscall_nr, enabled_exit_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

	event = trace_current_buffer_lock_reserve(&buffer,
			sys_data->exit_event->id, sizeof(*entry), 0, 0);
	if (!event)
		return;

	entry = ring_buffer_event_data(event);
	entry->nr = syscall_nr;
	entry->ret = syscall_get_return_value(current, regs);

	if (!filter_current_check_discard(buffer, sys_data->exit_event,
					  entry, event))
		trace_current_buffer_unlock_commit(buffer, event, 0, 0);
}

int reg_event_syscall_enter(struct ftrace_event_call *call)
{
	int ret = 0;
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;
	if (num < 0 || num >= NR_syscalls)
		return -ENOSYS;
	mutex_lock(&syscall_trace_lock);
	if (!sys_refcount_enter)
		ret = register_trace_sys_enter(ftrace_syscall_enter);
	if (!ret) {
		set_bit(num, enabled_enter_syscalls);
		sys_refcount_enter++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void unreg_event_syscall_enter(struct ftrace_event_call *call)
{
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;
	if (num < 0 || num >= NR_syscalls)
		return;
	mutex_lock(&syscall_trace_lock);
	sys_refcount_enter--;
	clear_bit(num, enabled_enter_syscalls);
	if (!sys_refcount_enter)
		unregister_trace_sys_enter(ftrace_syscall_enter);
	mutex_unlock(&syscall_trace_lock);
}

int reg_event_syscall_exit(struct ftrace_event_call *call)
{
	int ret = 0;
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;
	if (num < 0 || num >= NR_syscalls)
		return -ENOSYS;
	mutex_lock(&syscall_trace_lock);
	if (!sys_refcount_exit)
		ret = register_trace_sys_exit(ftrace_syscall_exit);
	if (!ret) {
		set_bit(num, enabled_exit_syscalls);
		sys_refcount_exit++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void unreg_event_syscall_exit(struct ftrace_event_call *call)
{
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;
	if (num < 0 || num >= NR_syscalls)
		return;
	mutex_lock(&syscall_trace_lock);
	sys_refcount_exit--;
	clear_bit(num, enabled_exit_syscalls);
	if (!sys_refcount_exit)
		unregister_trace_sys_exit(ftrace_syscall_exit);
	mutex_unlock(&syscall_trace_lock);
}

int init_syscall_trace(struct ftrace_event_call *call)
{
	int id;

	if (set_syscall_print_fmt(call) < 0)
		return -ENOMEM;

	id = trace_event_raw_init(call);

	if (id < 0) {
		free_syscall_print_fmt(call);
		return id;
	}

	return id;
}

unsigned long __init arch_syscall_addr(int nr)
{
	return (unsigned long)sys_call_table[nr];
}

int __init init_ftrace_syscalls(void)
{
	struct syscall_metadata *meta;
	unsigned long addr;
	int i;

	syscalls_metadata = kzalloc(sizeof(*syscalls_metadata) *
					NR_syscalls, GFP_KERNEL);
	if (!syscalls_metadata) {
		WARN_ON(1);
		return -ENOMEM;
	}

	for (i = 0; i < NR_syscalls; i++) {
		addr = arch_syscall_addr(i);
		meta = find_syscall_meta(addr);
		if (!meta)
			continue;

		meta->syscall_nr = i;
		syscalls_metadata[i] = meta;
	}

	return 0;
}
core_initcall(init_ftrace_syscalls);

#ifdef CONFIG_PERF_EVENTS

static DECLARE_BITMAP(enabled_perf_enter_syscalls, NR_syscalls);
static DECLARE_BITMAP(enabled_perf_exit_syscalls, NR_syscalls);
static int sys_perf_refcount_enter;
static int sys_perf_refcount_exit;

static void perf_syscall_enter(struct pt_regs *regs, long id)
{
	struct syscall_metadata *sys_data;
	struct syscall_trace_enter *rec;
	unsigned long flags;
	int syscall_nr;
	int rctx;
	int size;

	syscall_nr = syscall_get_nr(current, regs);
	if (!test_bit(syscall_nr, enabled_perf_enter_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

	/* get the size after alignment with the u32 buffer size field */
	size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec);
	size = ALIGN(size + sizeof(u32), sizeof(u64));
	size -= sizeof(u32);

	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
		      "perf buffer not large enough"))
		return;

	rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size,
				sys_data->enter_event->id, &rctx, &flags);
	if (!rec)
		return;

	rec->nr = syscall_nr;
	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
			       (unsigned long *)&rec->args);
	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
}

int perf_sysenter_enable(struct ftrace_event_call *call)
{
	int ret = 0;
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;

	mutex_lock(&syscall_trace_lock);
	if (!sys_perf_refcount_enter)
		ret = register_trace_sys_enter(perf_syscall_enter);
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall entry trace point");
	} else {
		set_bit(num, enabled_perf_enter_syscalls);
		sys_perf_refcount_enter++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void perf_sysenter_disable(struct ftrace_event_call *call)
{
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;

	mutex_lock(&syscall_trace_lock);
	sys_perf_refcount_enter--;
	clear_bit(num, enabled_perf_enter_syscalls);
	if (!sys_perf_refcount_enter)
		unregister_trace_sys_enter(perf_syscall_enter);
	mutex_unlock(&syscall_trace_lock);
}

static void perf_syscall_exit(struct pt_regs *regs, long ret)
{
	struct syscall_metadata *sys_data;
	struct syscall_trace_exit *rec;
	unsigned long flags;
	int syscall_nr;
	int rctx;
	int size;

	syscall_nr = syscall_get_nr(current, regs);
	if (!test_bit(syscall_nr, enabled_perf_exit_syscalls))
		return;

	sys_data = syscall_nr_to_meta(syscall_nr);
	if (!sys_data)
		return;

	/* We can probably do that at build time */
	size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
	size -= sizeof(u32);

	/*
	 * Impossible, but be paranoid with the future
	 * How to put this check outside runtime?
	 */
	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
		"exit event has grown above perf buffer size"))
		return;

	rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size,
				sys_data->exit_event->id, &rctx, &flags);
	if (!rec)
		return;

	rec->nr = syscall_nr;
	rec->ret = syscall_get_return_value(current, regs);

	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
}

int perf_sysexit_enable(struct ftrace_event_call *call)
{
	int ret = 0;
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;

	mutex_lock(&syscall_trace_lock);
	if (!sys_perf_refcount_exit)
		ret = register_trace_sys_exit(perf_syscall_exit);
	if (ret) {
		pr_info("event trace: Could not activate"
				"syscall exit trace point");
	} else {
		set_bit(num, enabled_perf_exit_syscalls);
		sys_perf_refcount_exit++;
	}
	mutex_unlock(&syscall_trace_lock);
	return ret;
}

void perf_sysexit_disable(struct ftrace_event_call *call)
{
	int num;

	num = ((struct syscall_metadata *)call->data)->syscall_nr;

	mutex_lock(&syscall_trace_lock);
	sys_perf_refcount_exit--;
	clear_bit(num, enabled_perf_exit_syscalls);
	if (!sys_perf_refcount_exit)
		unregister_trace_sys_exit(perf_syscall_exit);
	mutex_unlock(&syscall_trace_lock);
}

#endif /* CONFIG_PERF_EVENTS */
Commit	Line	Data
47788c58	1	#include <trace/syscall.h>
1c569f02	2	#include <trace/events/syscalls.h>
5a0e3ad6	3	#include <linux/slab.h>
ee08c6ec	4	#include <linux/kernel.h>
fb34a08c	5	#include <linux/ftrace.h>
cdd6c482	6	#include <linux/perf_event.h>
ee08c6ec FW	7	#include <asm/syscall.h>
	8
	9	#include "trace_output.h"
	10	#include "trace.h"
	11
5be71b61	12	static DEFINE_MUTEX(syscall_trace_lock);
fb34a08c JB	13	static int sys_refcount_enter;
fb34a08c JB	14	static int sys_refcount_exit;
57421dbb JB	15	static DECLARE_BITMAP(enabled_enter_syscalls, NR_syscalls);
57421dbb JB	16	static DECLARE_BITMAP(enabled_exit_syscalls, NR_syscalls);
ee08c6ec	17
c44fc770 FW	18	extern unsigned long __start_syscalls_metadata[];
	19	extern unsigned long __stop_syscalls_metadata[];
	20
	21	static struct syscall_metadata **syscalls_metadata;
	22
	23	static struct syscall_metadata *find_syscall_meta(unsigned long syscall)
	24	{
	25	struct syscall_metadata *start;
	26	struct syscall_metadata *stop;
	27	char str[KSYM_SYMBOL_LEN];
	28
	29
	30	start = (struct syscall_metadata *)__start_syscalls_metadata;
	31	stop = (struct syscall_metadata *)__stop_syscalls_metadata;
	32	kallsyms_lookup(syscall, NULL, NULL, NULL, str);
	33
	34	for ( ; start < stop; start++) {
	35	/*
	36	* Only compare after the "sys" prefix. Archs that use
	37	* syscall wrappers may have syscalls symbols aliases prefixed
	38	* with "SyS" instead of "sys", leading to an unwanted
	39	* mismatch.
	40	*/
	41	if (start->name && !strcmp(start->name + 3, str + 3))
	42	return start;
	43	}
	44	return NULL;
	45	}
	46
	47	static struct syscall_metadata *syscall_nr_to_meta(int nr)
	48	{
	49	if (!syscalls_metadata \|\| nr >= NR_syscalls \|\| nr < 0)
	50	return NULL;
	51
	52	return syscalls_metadata[nr];
	53	}
	54
bed1ffca FW	55	enum print_line_t
	56	print_syscall_enter(struct trace_iterator *iter, int flags)
	57	{
	58	struct trace_seq *s = &iter->seq;
	59	struct trace_entry *ent = iter->ent;
	60	struct syscall_trace_enter *trace;
	61	struct syscall_metadata *entry;
	62	int i, ret, syscall;
	63
64c12e04	64	trace = (typeof(trace))ent;
bed1ffca	65	syscall = trace->nr;
bed1ffca	66	entry = syscall_nr_to_meta(syscall);
64c12e04	67
bed1ffca FW	68	if (!entry)
	69	goto end;
	70
fcc19438	71	if (entry->enter_event->id != ent->type) {
64c12e04 JB	72	WARN_ON_ONCE(1);
	73	goto end;
	74	}
	75
bed1ffca FW	76	ret = trace_seq_printf(s, "%s(", entry->name);
	77	if (!ret)
	78	return TRACE_TYPE_PARTIAL_LINE;
	79
	80	for (i = 0; i < entry->nb_args; i++) {
	81	/* parameter types */
ba8b3a40	82	if (trace_flags & TRACE_ITER_VERBOSE) {
bed1ffca FW	83	ret = trace_seq_printf(s, "%s ", entry->types[i]);
	84	if (!ret)
	85	return TRACE_TYPE_PARTIAL_LINE;
	86	}
	87	/* parameter values */
4539f077	88	ret = trace_seq_printf(s, "%s: %lx%s", entry->args[i],
bed1ffca	89	trace->args[i],
4539f077	90	i == entry->nb_args - 1 ? "" : ", ");
bed1ffca FW	91	if (!ret)
	92	return TRACE_TYPE_PARTIAL_LINE;
	93	}
	94
4539f077 LZ	95	ret = trace_seq_putc(s, ')');
	96	if (!ret)
	97	return TRACE_TYPE_PARTIAL_LINE;
	98
bed1ffca	99	end:
4539f077 LZ	100	ret = trace_seq_putc(s, '\n');
	101	if (!ret)
	102	return TRACE_TYPE_PARTIAL_LINE;
	103
bed1ffca FW	104	return TRACE_TYPE_HANDLED;
	105	}
	106
	107	enum print_line_t
	108	print_syscall_exit(struct trace_iterator *iter, int flags)
	109	{
	110	struct trace_seq *s = &iter->seq;
	111	struct trace_entry *ent = iter->ent;
	112	struct syscall_trace_exit *trace;
	113	int syscall;
	114	struct syscall_metadata *entry;
	115	int ret;
	116
64c12e04	117	trace = (typeof(trace))ent;
bed1ffca	118	syscall = trace->nr;
bed1ffca	119	entry = syscall_nr_to_meta(syscall);
64c12e04	120
bed1ffca FW	121	if (!entry) {
	122	trace_seq_printf(s, "\n");
	123	return TRACE_TYPE_HANDLED;
	124	}
	125
fcc19438	126	if (entry->exit_event->id != ent->type) {
64c12e04 JB	127	WARN_ON_ONCE(1);
	128	return TRACE_TYPE_UNHANDLED;
	129	}
	130
bed1ffca FW	131	ret = trace_seq_printf(s, "%s -> 0x%lx\n", entry->name,
	132	trace->ret);
	133	if (!ret)
	134	return TRACE_TYPE_PARTIAL_LINE;
	135
	136	return TRACE_TYPE_HANDLED;
	137	}
	138
e6971969 LZ	139	extern char *__bad_type_size(void);
	140
	141	#define SYSCALL_FIELD(type, name) \
	142	sizeof(type) != sizeof(trace.name) ? \
	143	__bad_type_size() : \
26a50744 TZ	144	#type, #name, offsetof(typeof(trace), name), \
26a50744 TZ	145	sizeof(trace.name), is_signed_type(type)
e6971969	146
50307a45 LJ	147	static
	148	int __set_enter_print_fmt(struct syscall_metadata entry, char buf, int len)
	149	{
	150	int i;
	151	int pos = 0;
	152
	153	/* When len=0, we just calculate the needed length */
	154	#define LEN_OR_ZERO (len ? len - pos : 0)
	155
	156	pos += snprintf(buf + pos, LEN_OR_ZERO, "\"");
	157	for (i = 0; i < entry->nb_args; i++) {
	158	pos += snprintf(buf + pos, LEN_OR_ZERO, "%s: 0x%%0%zulx%s",
	159	entry->args[i], sizeof(unsigned long),
	160	i == entry->nb_args - 1 ? "" : ", ");
	161	}
	162	pos += snprintf(buf + pos, LEN_OR_ZERO, "\"");
	163
	164	for (i = 0; i < entry->nb_args; i++) {
	165	pos += snprintf(buf + pos, LEN_OR_ZERO,
	166	", ((unsigned long)(REC->%s))", entry->args[i]);
	167	}
	168
	169	#undef LEN_OR_ZERO
	170
	171	/* return the length of print_fmt */
	172	return pos;
	173	}
	174
	175	static int set_syscall_print_fmt(struct ftrace_event_call *call)
	176	{
	177	char *print_fmt;
	178	int len;
	179	struct syscall_metadata *entry = call->data;
	180
	181	if (entry->enter_event != call) {
	182	call->print_fmt = "\"0x%lx\", REC->ret";
	183	return 0;
	184	}
	185
	186	/* First: called with 0 length to calculate the needed length */
	187	len = __set_enter_print_fmt(entry, NULL, 0);
	188
	189	print_fmt = kmalloc(len + 1, GFP_KERNEL);
	190	if (!print_fmt)
	191	return -ENOMEM;
	192
	193	/* Second: actually write the @print_fmt */
	194	__set_enter_print_fmt(entry, print_fmt, len + 1);
	195	call->print_fmt = print_fmt;
	196
	197	return 0;
	198	}
	199
	200	static void free_syscall_print_fmt(struct ftrace_event_call *call)
	201	{
	202	struct syscall_metadata *entry = call->data;
	203
	204	if (entry->enter_event == call)
	205	kfree(call->print_fmt);
	206	}
	207
540b7b8d LZ	208	int syscall_enter_define_fields(struct ftrace_event_call *call)
	209	{
	210	struct syscall_trace_enter trace;
31c16b13	211	struct syscall_metadata *meta = call->data;
540b7b8d	212	int ret;
540b7b8d LZ	213	int i;
	214	int offset = offsetof(typeof(trace), args);
	215
0f1ef51d LJ	216	ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
	217	if (ret)
	218	return ret;
	219
540b7b8d	220	for (i = 0; i < meta->nb_args; i++) {
aeaeae11 FW	221	ret = trace_define_field(call, meta->types[i],
aeaeae11 FW	222	meta->args[i], offset,
43b51ead LZ	223	sizeof(unsigned long), 0,
43b51ead LZ	224	FILTER_OTHER);
540b7b8d LZ	225	offset += sizeof(unsigned long);
	226	}
	227
	228	return ret;
	229	}
	230
	231	int syscall_exit_define_fields(struct ftrace_event_call *call)
	232	{
	233	struct syscall_trace_exit trace;
	234	int ret;
	235
0f1ef51d LJ	236	ret = trace_define_field(call, SYSCALL_FIELD(int, nr), FILTER_OTHER);
	237	if (ret)
	238	return ret;
	239
26a50744	240	ret = trace_define_field(call, SYSCALL_FIELD(long, ret),
43b51ead	241	FILTER_OTHER);
540b7b8d LZ	242
	243	return ret;
	244	}
	245
fb34a08c	246	void ftrace_syscall_enter(struct pt_regs *regs, long id)
ee08c6ec	247	{
bed1ffca FW	248	struct syscall_trace_enter *entry;
	249	struct syscall_metadata *sys_data;
	250	struct ring_buffer_event *event;
e77405ad	251	struct ring_buffer *buffer;
bed1ffca	252	int size;
ee08c6ec FW	253	int syscall_nr;
	254
	255	syscall_nr = syscall_get_nr(current, regs);
cd0980fc HB	256	if (syscall_nr < 0)
cd0980fc HB	257	return;
fb34a08c JB	258	if (!test_bit(syscall_nr, enabled_enter_syscalls))
fb34a08c JB	259	return;
ee08c6ec	260
bed1ffca FW	261	sys_data = syscall_nr_to_meta(syscall_nr);
	262	if (!sys_data)
	263	return;
	264
	265	size = sizeof(entry) + sizeof(unsigned long) sys_data->nb_args;
	266
fcc19438 LJ	267	event = trace_current_buffer_lock_reserve(&buffer,
fcc19438 LJ	268	sys_data->enter_event->id, size, 0, 0);
bed1ffca FW	269	if (!event)
	270	return;
	271
	272	entry = ring_buffer_event_data(event);
	273	entry->nr = syscall_nr;
	274	syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args);
	275
e77405ad SR	276	if (!filter_current_check_discard(buffer, sys_data->enter_event,
	277	entry, event))
	278	trace_current_buffer_unlock_commit(buffer, event, 0, 0);
ee08c6ec FW	279	}
ee08c6ec FW	280
fb34a08c	281	void ftrace_syscall_exit(struct pt_regs *regs, long ret)
ee08c6ec	282	{
bed1ffca FW	283	struct syscall_trace_exit *entry;
	284	struct syscall_metadata *sys_data;
	285	struct ring_buffer_event *event;
e77405ad	286	struct ring_buffer *buffer;
ee08c6ec FW	287	int syscall_nr;
	288
	289	syscall_nr = syscall_get_nr(current, regs);
cd0980fc HB	290	if (syscall_nr < 0)
cd0980fc HB	291	return;
fb34a08c JB	292	if (!test_bit(syscall_nr, enabled_exit_syscalls))
fb34a08c JB	293	return;
ee08c6ec	294
bed1ffca FW	295	sys_data = syscall_nr_to_meta(syscall_nr);
	296	if (!sys_data)
	297	return;
	298
fcc19438 LJ	299	event = trace_current_buffer_lock_reserve(&buffer,
fcc19438 LJ	300	sys_data->exit_event->id, sizeof(*entry), 0, 0);
bed1ffca FW	301	if (!event)
	302	return;
	303
	304	entry = ring_buffer_event_data(event);
	305	entry->nr = syscall_nr;
	306	entry->ret = syscall_get_return_value(current, regs);
	307
e77405ad SR	308	if (!filter_current_check_discard(buffer, sys_data->exit_event,
	309	entry, event))
	310	trace_current_buffer_unlock_commit(buffer, event, 0, 0);
ee08c6ec FW	311	}
ee08c6ec FW	312
bd1a5c84	313	int reg_event_syscall_enter(struct ftrace_event_call *call)
ee08c6ec	314	{
fb34a08c JB	315	int ret = 0;
fb34a08c JB	316	int num;
fb34a08c	317
c252f657	318	num = ((struct syscall_metadata *)call->data)->syscall_nr;
57421dbb	319	if (num < 0 \|\| num >= NR_syscalls)
fb34a08c JB	320	return -ENOSYS;
	321	mutex_lock(&syscall_trace_lock);
	322	if (!sys_refcount_enter)
1c569f02	323	ret = register_trace_sys_enter(ftrace_syscall_enter);
3b8e4273	324	if (!ret) {
fb34a08c JB	325	set_bit(num, enabled_enter_syscalls);
	326	sys_refcount_enter++;
	327	}
	328	mutex_unlock(&syscall_trace_lock);
	329	return ret;
ee08c6ec FW	330	}
ee08c6ec FW	331
bd1a5c84	332	void unreg_event_syscall_enter(struct ftrace_event_call *call)
ee08c6ec	333	{
fb34a08c	334	int num;
ee08c6ec	335
c252f657	336	num = ((struct syscall_metadata *)call->data)->syscall_nr;
57421dbb	337	if (num < 0 \|\| num >= NR_syscalls)
fb34a08c JB	338	return;
	339	mutex_lock(&syscall_trace_lock);
	340	sys_refcount_enter--;
	341	clear_bit(num, enabled_enter_syscalls);
	342	if (!sys_refcount_enter)
1c569f02	343	unregister_trace_sys_enter(ftrace_syscall_enter);
fb34a08c JB	344	mutex_unlock(&syscall_trace_lock);
fb34a08c JB	345	}
ee08c6ec	346
bd1a5c84	347	int reg_event_syscall_exit(struct ftrace_event_call *call)
ee08c6ec	348	{
fb34a08c JB	349	int ret = 0;
fb34a08c JB	350	int num;
fb34a08c	351
c252f657	352	num = ((struct syscall_metadata *)call->data)->syscall_nr;
57421dbb	353	if (num < 0 \|\| num >= NR_syscalls)
fb34a08c JB	354	return -ENOSYS;
	355	mutex_lock(&syscall_trace_lock);
	356	if (!sys_refcount_exit)
1c569f02	357	ret = register_trace_sys_exit(ftrace_syscall_exit);
3b8e4273	358	if (!ret) {
fb34a08c JB	359	set_bit(num, enabled_exit_syscalls);
fb34a08c JB	360	sys_refcount_exit++;
ee08c6ec	361	}
fb34a08c JB	362	mutex_unlock(&syscall_trace_lock);
	363	return ret;
	364	}
ee08c6ec	365
bd1a5c84	366	void unreg_event_syscall_exit(struct ftrace_event_call *call)
fb34a08c JB	367	{
fb34a08c JB	368	int num;
ee08c6ec	369
c252f657	370	num = ((struct syscall_metadata *)call->data)->syscall_nr;
57421dbb	371	if (num < 0 \|\| num >= NR_syscalls)
fb34a08c JB	372	return;
	373	mutex_lock(&syscall_trace_lock);
	374	sys_refcount_exit--;
	375	clear_bit(num, enabled_exit_syscalls);
	376	if (!sys_refcount_exit)
1c569f02	377	unregister_trace_sys_exit(ftrace_syscall_exit);
fb34a08c	378	mutex_unlock(&syscall_trace_lock);
ee08c6ec	379	}
fb34a08c	380
a1301da0 LJ	381	int init_syscall_trace(struct ftrace_event_call *call)
	382	{
	383	int id;
	384
50307a45 LJ	385	if (set_syscall_print_fmt(call) < 0)
	386	return -ENOMEM;
	387
c7ef3a90 SR	388	id = trace_event_raw_init(call);
	389
	390	if (id < 0) {
50307a45	391	free_syscall_print_fmt(call);
c7ef3a90	392	return id;
50307a45	393	}
c7ef3a90 SR	394
c7ef3a90 SR	395	return id;
a1301da0 LJ	396	}
a1301da0 LJ	397
e7b8e675 MF	398	unsigned long __init arch_syscall_addr(int nr)
	399	{
	400	return (unsigned long)sys_call_table[nr];
	401	}
	402
c44fc770 FW	403	int __init init_ftrace_syscalls(void)
	404	{
	405	struct syscall_metadata *meta;
	406	unsigned long addr;
	407	int i;
	408
	409	syscalls_metadata = kzalloc(sizeof(syscalls_metadata)
	410	NR_syscalls, GFP_KERNEL);
	411	if (!syscalls_metadata) {
	412	WARN_ON(1);
	413	return -ENOMEM;
	414	}
	415
	416	for (i = 0; i < NR_syscalls; i++) {
	417	addr = arch_syscall_addr(i);
	418	meta = find_syscall_meta(addr);
c252f657 LJ	419	if (!meta)
	420	continue;
	421
	422	meta->syscall_nr = i;
c44fc770 FW	423	syscalls_metadata[i] = meta;
	424	}
	425
	426	return 0;
	427	}
	428	core_initcall(init_ftrace_syscalls);
	429
07b139c8	430	#ifdef CONFIG_PERF_EVENTS
19007a67	431
97d5a220 FW	432	static DECLARE_BITMAP(enabled_perf_enter_syscalls, NR_syscalls);
	433	static DECLARE_BITMAP(enabled_perf_exit_syscalls, NR_syscalls);
	434	static int sys_perf_refcount_enter;
	435	static int sys_perf_refcount_exit;
f4b5ffcc	436
97d5a220	437	static void perf_syscall_enter(struct pt_regs *regs, long id)
f4b5ffcc JB	438	{
f4b5ffcc JB	439	struct syscall_metadata *sys_data;
20ab4425 FW	440	struct syscall_trace_enter *rec;
20ab4425 FW	441	unsigned long flags;
f4b5ffcc	442	int syscall_nr;
4ed7c92d	443	int rctx;
19007a67	444	int size;
f4b5ffcc JB	445
f4b5ffcc JB	446	syscall_nr = syscall_get_nr(current, regs);
97d5a220	447	if (!test_bit(syscall_nr, enabled_perf_enter_syscalls))
f4b5ffcc JB	448	return;
	449
	450	sys_data = syscall_nr_to_meta(syscall_nr);
	451	if (!sys_data)
	452	return;
	453
19007a67 FW	454	/* get the size after alignment with the u32 buffer size field */
	455	size = sizeof(unsigned long) * sys_data->nb_args + sizeof(*rec);
	456	size = ALIGN(size + sizeof(u32), sizeof(u64));
	457	size -= sizeof(u32);
	458
97d5a220 FW	459	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
97d5a220 FW	460	"perf buffer not large enough"))
20ab4425 FW	461	return;
20ab4425 FW	462
97d5a220	463	rec = (struct syscall_trace_enter *)perf_trace_buf_prepare(size,
430ad5a6 XG	464	sys_data->enter_event->id, &rctx, &flags);
	465	if (!rec)
	466	return;
20ab4425	467
20ab4425 FW	468	rec->nr = syscall_nr;
	469	syscall_get_arguments(current, regs, 0, sys_data->nb_args,
	470	(unsigned long *)&rec->args);
97d5a220	471	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
f4b5ffcc JB	472	}
f4b5ffcc JB	473
97d5a220	474	int perf_sysenter_enable(struct ftrace_event_call *call)
f4b5ffcc JB	475	{
	476	int ret = 0;
	477	int num;
	478
3bbe84e9	479	num = ((struct syscall_metadata *)call->data)->syscall_nr;
f4b5ffcc JB	480
f4b5ffcc JB	481	mutex_lock(&syscall_trace_lock);
97d5a220 FW	482	if (!sys_perf_refcount_enter)
97d5a220 FW	483	ret = register_trace_sys_enter(perf_syscall_enter);
f4b5ffcc JB	484	if (ret) {
	485	pr_info("event trace: Could not activate"
	486	"syscall entry trace point");
	487	} else {
97d5a220 FW	488	set_bit(num, enabled_perf_enter_syscalls);
97d5a220 FW	489	sys_perf_refcount_enter++;
f4b5ffcc JB	490	}
	491	mutex_unlock(&syscall_trace_lock);
	492	return ret;
	493	}
	494
97d5a220	495	void perf_sysenter_disable(struct ftrace_event_call *call)
f4b5ffcc JB	496	{
	497	int num;
	498
3bbe84e9	499	num = ((struct syscall_metadata *)call->data)->syscall_nr;
f4b5ffcc JB	500
f4b5ffcc JB	501	mutex_lock(&syscall_trace_lock);
97d5a220 FW	502	sys_perf_refcount_enter--;
	503	clear_bit(num, enabled_perf_enter_syscalls);
	504	if (!sys_perf_refcount_enter)
	505	unregister_trace_sys_enter(perf_syscall_enter);
f4b5ffcc JB	506	mutex_unlock(&syscall_trace_lock);
	507	}
	508
97d5a220	509	static void perf_syscall_exit(struct pt_regs *regs, long ret)
f4b5ffcc JB	510	{
f4b5ffcc JB	511	struct syscall_metadata *sys_data;
20ab4425 FW	512	struct syscall_trace_exit *rec;
20ab4425 FW	513	unsigned long flags;
f4b5ffcc	514	int syscall_nr;
4ed7c92d	515	int rctx;
20ab4425	516	int size;
f4b5ffcc JB	517
f4b5ffcc JB	518	syscall_nr = syscall_get_nr(current, regs);
97d5a220	519	if (!test_bit(syscall_nr, enabled_perf_exit_syscalls))
f4b5ffcc JB	520	return;
	521
	522	sys_data = syscall_nr_to_meta(syscall_nr);
	523	if (!sys_data)
	524	return;
	525
20ab4425 FW	526	/* We can probably do that at build time */
	527	size = ALIGN(sizeof(*rec) + sizeof(u32), sizeof(u64));
	528	size -= sizeof(u32);
19007a67	529
20ab4425 FW	530	/*
	531	* Impossible, but be paranoid with the future
	532	* How to put this check outside runtime?
	533	*/
97d5a220 FW	534	if (WARN_ONCE(size > PERF_MAX_TRACE_SIZE,
97d5a220 FW	535	"exit event has grown above perf buffer size"))
20ab4425 FW	536	return;
20ab4425 FW	537
97d5a220	538	rec = (struct syscall_trace_exit *)perf_trace_buf_prepare(size,
430ad5a6 XG	539	sys_data->exit_event->id, &rctx, &flags);
	540	if (!rec)
	541	return;
20ab4425	542
20ab4425 FW	543	rec->nr = syscall_nr;
	544	rec->ret = syscall_get_return_value(current, regs);
	545
97d5a220	546	perf_trace_buf_submit(rec, size, rctx, 0, 1, flags, regs);
f4b5ffcc JB	547	}
f4b5ffcc JB	548
97d5a220	549	int perf_sysexit_enable(struct ftrace_event_call *call)
f4b5ffcc JB	550	{
	551	int ret = 0;
	552	int num;
	553
3bbe84e9	554	num = ((struct syscall_metadata *)call->data)->syscall_nr;
f4b5ffcc JB	555
f4b5ffcc JB	556	mutex_lock(&syscall_trace_lock);
97d5a220 FW	557	if (!sys_perf_refcount_exit)
97d5a220 FW	558	ret = register_trace_sys_exit(perf_syscall_exit);
f4b5ffcc JB	559	if (ret) {
f4b5ffcc JB	560	pr_info("event trace: Could not activate"
6574658b	561	"syscall exit trace point");
f4b5ffcc	562	} else {
97d5a220 FW	563	set_bit(num, enabled_perf_exit_syscalls);
97d5a220 FW	564	sys_perf_refcount_exit++;
f4b5ffcc JB	565	}
	566	mutex_unlock(&syscall_trace_lock);
	567	return ret;
	568	}
	569
97d5a220	570	void perf_sysexit_disable(struct ftrace_event_call *call)
f4b5ffcc JB	571	{
	572	int num;
	573
3bbe84e9	574	num = ((struct syscall_metadata *)call->data)->syscall_nr;
f4b5ffcc JB	575
f4b5ffcc JB	576	mutex_lock(&syscall_trace_lock);
97d5a220 FW	577	sys_perf_refcount_exit--;
	578	clear_bit(num, enabled_perf_exit_syscalls);
	579	if (!sys_perf_refcount_exit)
	580	unregister_trace_sys_exit(perf_syscall_exit);
f4b5ffcc JB	581	mutex_unlock(&syscall_trace_lock);
	582	}
	583
07b139c8	584	#endif /* CONFIG_PERF_EVENTS */
f4b5ffcc	585