[deliverable/linux.git] / tools / perf / builtin-trace.c

#include "builtin.h"
#include "util/color.h"
#include "util/evlist.h"
#include "util/machine.h"
#include "util/thread.h"
#include "util/parse-options.h"
#include "util/thread_map.h"
#include "event-parse.h"

#include <libaudit.h>
#include <stdlib.h>

static struct syscall_fmt {
	const char *name;
	const char *alias;
	bool	   errmsg;
	bool	   timeout;
} syscall_fmts[] = {
	{ .name	    = "arch_prctl", .errmsg = true, .alias = "prctl", },
	{ .name	    = "fstat",	    .errmsg = true, .alias = "newfstat", },
	{ .name	    = "fstatat",    .errmsg = true, .alias = "newfstatat", },
	{ .name	    = "futex",	    .errmsg = true, },
	{ .name	    = "poll",	    .errmsg = true, .timeout = true, },
	{ .name	    = "ppoll",	    .errmsg = true, .timeout = true, },
	{ .name	    = "read",	    .errmsg = true, },
	{ .name	    = "recvfrom",   .errmsg = true, },
	{ .name	    = "select",	    .errmsg = true, .timeout = true, },
	{ .name	    = "stat",	    .errmsg = true, .alias = "newstat", },
};

static int syscall_fmt__cmp(const void *name, const void *fmtp)
{
	const struct syscall_fmt *fmt = fmtp;
	return strcmp(name, fmt->name);
}

static struct syscall_fmt *syscall_fmt__find(const char *name)
{
	const int nmemb = ARRAY_SIZE(syscall_fmts);
	return bsearch(name, syscall_fmts, nmemb, sizeof(struct syscall_fmt), syscall_fmt__cmp);
}

struct syscall {
	struct event_format *tp_format;
	const char	    *name;
	struct syscall_fmt  *fmt;
};

struct thread_trace {
	u64		  entry_time;
	u64		  exit_time;
	bool		  entry_pending;
	char		  *entry_str;
};

static struct thread_trace *thread_trace__new(void)
{
	return zalloc(sizeof(struct thread_trace));
}

static struct thread_trace *thread__trace(struct thread *thread)
{
	if (thread == NULL)
		goto fail;

	if (thread->priv == NULL)
		thread->priv = thread_trace__new();

	if (thread->priv == NULL)
		goto fail;

	return thread->priv;
fail:
	color_fprintf(stdout, PERF_COLOR_RED,
		      "WARNING: not enough memory, dropping samples!\n");
	return NULL;
}

struct trace {
	int			audit_machine;
	struct {
		int		max;
		struct syscall  *table;
	} syscalls;
	struct perf_record_opts opts;
	struct machine		host;
	u64			base_time;
	bool			multiple_threads;
};

static size_t trace__fprintf_tstamp(struct trace *trace, u64 tstamp, FILE *fp)
{
	double ts = (double)(tstamp - trace->base_time) / NSEC_PER_MSEC;

	return fprintf(fp, "%10.3f: ", ts);
}

static bool done = false;

static void sig_handler(int sig __maybe_unused)
{
	done = true;
}

static size_t trace__fprintf_entry_head(struct trace *trace, struct thread *thread,
					u64 tstamp, FILE *fp)
{
	size_t printed = trace__fprintf_tstamp(trace, tstamp, fp);

	if (trace->multiple_threads)
		printed += fprintf(fp, "%d ", thread->pid);

	return printed;
}

static int trace__process_event(struct machine *machine, union perf_event *event)
{
	int ret = 0;

	switch (event->header.type) {
	case PERF_RECORD_LOST:
		color_fprintf(stdout, PERF_COLOR_RED,
			      "LOST %" PRIu64 " events!\n", event->lost.lost);
		ret = machine__process_lost_event(machine, event);
	default:
		ret = machine__process_event(machine, event);
		break;
	}

	return ret;
}

static int trace__tool_process(struct perf_tool *tool __maybe_unused,
			       union perf_event *event,
			       struct perf_sample *sample __maybe_unused,
			       struct machine *machine)
{
	return trace__process_event(machine, event);
}

static int trace__symbols_init(struct trace *trace, struct perf_evlist *evlist)
{
	int err = symbol__init();

	if (err)
		return err;

	machine__init(&trace->host, "", HOST_KERNEL_ID);
	machine__create_kernel_maps(&trace->host);

	if (perf_target__has_task(&trace->opts.target)) {
		err = perf_event__synthesize_thread_map(NULL, evlist->threads,
							trace__tool_process,
							&trace->host);
	} else {
		err = perf_event__synthesize_threads(NULL, trace__tool_process,
						     &trace->host);
	}

	if (err)
		symbol__exit();

	return err;
}

static int trace__read_syscall_info(struct trace *trace, int id)
{
	char tp_name[128];
	struct syscall *sc;
	const char *name = audit_syscall_to_name(id, trace->audit_machine);

	if (name == NULL)
		return -1;

	if (id > trace->syscalls.max) {
		struct syscall *nsyscalls = realloc(trace->syscalls.table, (id + 1) * sizeof(*sc));

		if (nsyscalls == NULL)
			return -1;

		if (trace->syscalls.max != -1) {
			memset(nsyscalls + trace->syscalls.max + 1, 0,
			       (id - trace->syscalls.max) * sizeof(*sc));
		} else {
			memset(nsyscalls, 0, (id + 1) * sizeof(*sc));
		}

		trace->syscalls.table = nsyscalls;
		trace->syscalls.max   = id;
	}

	sc = trace->syscalls.table + id;
	sc->name = name;
	sc->fmt  = syscall_fmt__find(sc->name);

	snprintf(tp_name, sizeof(tp_name), "sys_enter_%s", sc->name);
	sc->tp_format = event_format__new("syscalls", tp_name);

	if (sc->tp_format == NULL && sc->fmt && sc->fmt->alias) {
		snprintf(tp_name, sizeof(tp_name), "sys_enter_%s", sc->fmt->alias);
		sc->tp_format = event_format__new("syscalls", tp_name);
	}

	return sc->tp_format != NULL ? 0 : -1;
}

static size_t syscall__scnprintf_args(struct syscall *sc, char *bf, size_t size,
				      unsigned long *args)
{
	int i = 0;
	size_t printed = 0;

	if (sc->tp_format != NULL) {
		struct format_field *field;

		for (field = sc->tp_format->format.fields->next; field; field = field->next) {
			printed += scnprintf(bf + printed, size - printed,
					     "%s%s: %ld", printed ? ", " : "",
					     field->name, args[i++]);
		}
	} else {
		while (i < 6) {
			printed += scnprintf(bf + printed, size - printed,
					     "%sarg%d: %ld",
					     printed ? ", " : "", i, args[i]);
			++i;
		}
	}

	return printed;
}

typedef int (*tracepoint_handler)(struct trace *trace, struct perf_evsel *evsel,
				  struct perf_sample *sample);

static struct syscall *trace__syscall_info(struct trace *trace,
					   struct perf_evsel *evsel,
					   struct perf_sample *sample)
{
	int id = perf_evsel__intval(evsel, sample, "id");

	if (id < 0) {
		printf("Invalid syscall %d id, skipping...\n", id);
		return NULL;
	}

	if ((id > trace->syscalls.max || trace->syscalls.table[id].name == NULL) &&
	    trace__read_syscall_info(trace, id))
		goto out_cant_read;

	if ((id > trace->syscalls.max || trace->syscalls.table[id].name == NULL))
		goto out_cant_read;

	return &trace->syscalls.table[id];

out_cant_read:
	printf("Problems reading syscall %d information\n", id);
	return NULL;
}

static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel,
			    struct perf_sample *sample)
{
	char *msg;
	void *args;
	size_t printed = 0;
	struct thread *thread = machine__findnew_thread(&trace->host, sample->tid);
	struct syscall *sc = trace__syscall_info(trace, evsel, sample);
	struct thread_trace *ttrace = thread__trace(thread);

	if (ttrace == NULL || sc == NULL)
		return -1;

	args = perf_evsel__rawptr(evsel, sample, "args");
	if (args == NULL) {
		printf("Problems reading syscall arguments\n");
		return -1;
	}

	ttrace = thread->priv;

	if (ttrace->entry_str == NULL) {
		ttrace->entry_str = malloc(1024);
		if (!ttrace->entry_str)
			return -1;
	}

	ttrace->entry_time = sample->time;
	msg = ttrace->entry_str;
	printed += scnprintf(msg + printed, 1024 - printed, "%s(", sc->name);

	printed += syscall__scnprintf_args(sc, msg + printed, 1024 - printed,  args);

	if (!strcmp(sc->name, "exit_group") || !strcmp(sc->name, "exit")) {
		trace__fprintf_entry_head(trace, thread, sample->time, stdout);
		printf("%-70s\n", ttrace->entry_str);
	} else
		ttrace->entry_pending = true;

	return 0;
}

static int trace__sys_exit(struct trace *trace, struct perf_evsel *evsel,
			   struct perf_sample *sample)
{
	int ret;
	struct thread *thread = machine__findnew_thread(&trace->host, sample->tid);
	struct thread_trace *ttrace = thread__trace(thread);
	struct syscall *sc = trace__syscall_info(trace, evsel, sample);

	if (ttrace == NULL || sc == NULL)
		return -1;

	ret = perf_evsel__intval(evsel, sample, "ret");

	ttrace = thread->priv;

	ttrace->exit_time = sample->time;

	trace__fprintf_entry_head(trace, thread, sample->time, stdout);

	if (ttrace->entry_pending) {
		printf("%-70s", ttrace->entry_str);
	} else {
		printf(" ... [");
		color_fprintf(stdout, PERF_COLOR_YELLOW, "continued");
		printf("]: %s()", sc->name);
	}

	if (ret < 0 && sc->fmt && sc->fmt->errmsg) {
		char bf[256];
		const char *emsg = strerror_r(-ret, bf, sizeof(bf)),
			   *e = audit_errno_to_name(-ret);

		printf(") = -1 %s %s", e, emsg);
	} else if (ret == 0 && sc->fmt && sc->fmt->timeout)
		printf(") = 0 Timeout");
	else
		printf(") = %d", ret);

	putchar('\n');

	ttrace->entry_pending = false;

	return 0;
}

static int trace__run(struct trace *trace, int argc, const char **argv)
{
	struct perf_evlist *evlist = perf_evlist__new(NULL, NULL);
	struct perf_evsel *evsel;
	int err = -1, i, nr_events = 0, before;
	const bool forks = argc > 0;

	if (evlist == NULL) {
		printf("Not enough memory to run!\n");
		goto out;
	}

	if (perf_evlist__add_newtp(evlist, "raw_syscalls", "sys_enter", trace__sys_enter) ||
	    perf_evlist__add_newtp(evlist, "raw_syscalls", "sys_exit", trace__sys_exit)) {
		printf("Couldn't read the raw_syscalls tracepoints information!\n");
		goto out_delete_evlist;
	}

	err = perf_evlist__create_maps(evlist, &trace->opts.target);
	if (err < 0) {
		printf("Problems parsing the target to trace, check your options!\n");
		goto out_delete_evlist;
	}

	err = trace__symbols_init(trace, evlist);
	if (err < 0) {
		printf("Problems initializing symbol libraries!\n");
		goto out_delete_evlist;
	}

	perf_evlist__config_attrs(evlist, &trace->opts);

	signal(SIGCHLD, sig_handler);
	signal(SIGINT, sig_handler);

	if (forks) {
		err = perf_evlist__prepare_workload(evlist, &trace->opts, argv);
		if (err < 0) {
			printf("Couldn't run the workload!\n");
			goto out_delete_evlist;
		}
	}

	err = perf_evlist__open(evlist);
	if (err < 0) {
		printf("Couldn't create the events: %s\n", strerror(errno));
		goto out_delete_evlist;
	}

	err = perf_evlist__mmap(evlist, UINT_MAX, false);
	if (err < 0) {
		printf("Couldn't mmap the events: %s\n", strerror(errno));
		goto out_delete_evlist;
	}

	perf_evlist__enable(evlist);

	if (forks)
		perf_evlist__start_workload(evlist);

	trace->multiple_threads = evlist->threads->map[0] == -1 || evlist->threads->nr > 1;
again:
	before = nr_events;

	for (i = 0; i < evlist->nr_mmaps; i++) {
		union perf_event *event;

		while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
			const u32 type = event->header.type;
			tracepoint_handler handler;
			struct perf_sample sample;

			++nr_events;

			err = perf_evlist__parse_sample(evlist, event, &sample);
			if (err) {
				printf("Can't parse sample, err = %d, skipping...\n", err);
				continue;
			}

			if (trace->base_time == 0)
				trace->base_time = sample.time;

			if (type != PERF_RECORD_SAMPLE) {
				trace__process_event(&trace->host, event);
				continue;
			}

			evsel = perf_evlist__id2evsel(evlist, sample.id);
			if (evsel == NULL) {
				printf("Unknown tp ID %" PRIu64 ", skipping...\n", sample.id);
				continue;
			}

			if (sample.raw_data == NULL) {
				printf("%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
				       perf_evsel__name(evsel), sample.tid,
				       sample.cpu, sample.raw_size);
				continue;
			}

			if (sample.raw_data == NULL) {
				printf("%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
				       perf_evsel__name(evsel), sample.tid,
				       sample.cpu, sample.raw_size);
				continue;
			}

			handler = evsel->handler.func;
			handler(trace, evsel, &sample);
		}
	}

	if (nr_events == before) {
		if (done)
			goto out_delete_evlist;

		poll(evlist->pollfd, evlist->nr_fds, -1);
	}

	if (done)
		perf_evlist__disable(evlist);

	goto again;

out_delete_evlist:
	perf_evlist__delete(evlist);
out:
	return err;
}

int cmd_trace(int argc, const char **argv, const char *prefix __maybe_unused)
{
	const char * const trace_usage[] = {
		"perf trace [<options>] [<command>]",
		"perf trace [<options>] -- <command> [<options>]",
		NULL
	};
	struct trace trace = {
		.audit_machine = audit_detect_machine(),
		.syscalls = {
			. max = -1,
		},
		.opts = {
			.target = {
				.uid	   = UINT_MAX,
				.uses_mmap = true,
			},
			.user_freq     = UINT_MAX,
			.user_interval = ULLONG_MAX,
			.no_delay      = true,
			.mmap_pages    = 1024,
		},
	};
	const struct option trace_options[] = {
	OPT_STRING('p', "pid", &trace.opts.target.pid, "pid",
		    "trace events on existing process id"),
	OPT_STRING(0, "tid", &trace.opts.target.tid, "tid",
		    "trace events on existing thread id"),
	OPT_BOOLEAN(0, "all-cpus", &trace.opts.target.system_wide,
		    "system-wide collection from all CPUs"),
	OPT_STRING(0, "cpu", &trace.opts.target.cpu_list, "cpu",
		    "list of cpus to monitor"),
	OPT_BOOLEAN(0, "no-inherit", &trace.opts.no_inherit,
		    "child tasks do not inherit counters"),
	OPT_UINTEGER(0, "mmap-pages", &trace.opts.mmap_pages,
		     "number of mmap data pages"),
	OPT_STRING(0, "uid", &trace.opts.target.uid_str, "user",
		   "user to profile"),
	OPT_END()
	};
	int err;
	char bf[BUFSIZ];

	argc = parse_options(argc, argv, trace_options, trace_usage, 0);

	err = perf_target__validate(&trace.opts.target);
	if (err) {
		perf_target__strerror(&trace.opts.target, err, bf, sizeof(bf));
		printf("%s", bf);
		return err;
	}

	err = perf_target__parse_uid(&trace.opts.target);
	if (err) {
		perf_target__strerror(&trace.opts.target, err, bf, sizeof(bf));
		printf("%s", bf);
		return err;
	}

	if (!argc && perf_target__none(&trace.opts.target))
		trace.opts.target.system_wide = true;

	return trace__run(&trace, argc, argv);
}
Commit	Line	Data
514f1c67	1	#include "builtin.h"
752fde44	2	#include "util/color.h"
514f1c67	3	#include "util/evlist.h"
752fde44 ACM	4	#include "util/machine.h"
752fde44 ACM	5	#include "util/thread.h"
514f1c67 ACM	6	#include "util/parse-options.h"
	7	#include "util/thread_map.h"
	8	#include "event-parse.h"
	9
	10	#include <libaudit.h>
	11	#include <stdlib.h>
	12
	13	static struct syscall_fmt {
	14	const char *name;
aec1930b	15	const char *alias;
514f1c67 ACM	16	bool errmsg;
	17	bool timeout;
	18	} syscall_fmts[] = {
aec1930b ACM	19	{ .name = "arch_prctl", .errmsg = true, .alias = "prctl", },
	20	{ .name = "fstat", .errmsg = true, .alias = "newfstat", },
	21	{ .name = "fstatat", .errmsg = true, .alias = "newfstatat", },
	22	{ .name = "futex", .errmsg = true, },
	23	{ .name = "poll", .errmsg = true, .timeout = true, },
	24	{ .name = "ppoll", .errmsg = true, .timeout = true, },
	25	{ .name = "read", .errmsg = true, },
	26	{ .name = "recvfrom", .errmsg = true, },
	27	{ .name = "select", .errmsg = true, .timeout = true, },
	28	{ .name = "stat", .errmsg = true, .alias = "newstat", },
514f1c67 ACM	29	};
	30
	31	static int syscall_fmt__cmp(const void name, const void fmtp)
	32	{
	33	const struct syscall_fmt *fmt = fmtp;
	34	return strcmp(name, fmt->name);
	35	}
	36
	37	static struct syscall_fmt syscall_fmt__find(const char name)
	38	{
	39	const int nmemb = ARRAY_SIZE(syscall_fmts);
	40	return bsearch(name, syscall_fmts, nmemb, sizeof(struct syscall_fmt), syscall_fmt__cmp);
	41	}
	42
	43	struct syscall {
	44	struct event_format *tp_format;
	45	const char *name;
	46	struct syscall_fmt *fmt;
	47	};
	48
752fde44 ACM	49	struct thread_trace {
	50	u64 entry_time;
	51	u64 exit_time;
	52	bool entry_pending;
	53	char *entry_str;
	54	};
	55
	56	static struct thread_trace *thread_trace__new(void)
	57	{
	58	return zalloc(sizeof(struct thread_trace));
	59	}
	60
	61	static struct thread_trace thread__trace(struct thread thread)
	62	{
	63	if (thread == NULL)
	64	goto fail;
	65
	66	if (thread->priv == NULL)
	67	thread->priv = thread_trace__new();
	68
	69	if (thread->priv == NULL)
	70	goto fail;
	71
	72	return thread->priv;
	73	fail:
	74	color_fprintf(stdout, PERF_COLOR_RED,
	75	"WARNING: not enough memory, dropping samples!\n");
	76	return NULL;
	77	}
	78
514f1c67 ACM	79	struct trace {
	80	int audit_machine;
	81	struct {
	82	int max;
	83	struct syscall *table;
	84	} syscalls;
	85	struct perf_record_opts opts;
752fde44 ACM	86	struct machine host;
	87	u64 base_time;
	88	bool multiple_threads;
514f1c67 ACM	89	};
514f1c67 ACM	90
752fde44 ACM	91	static size_t trace__fprintf_tstamp(struct trace trace, u64 tstamp, FILE fp)
	92	{
	93	double ts = (double)(tstamp - trace->base_time) / NSEC_PER_MSEC;
	94
	95	return fprintf(fp, "%10.3f: ", ts);
	96	}
	97
f15eb531 NK	98	static bool done = false;
	99
	100	static void sig_handler(int sig __maybe_unused)
	101	{
	102	done = true;
	103	}
	104
752fde44 ACM	105	static size_t trace__fprintf_entry_head(struct trace trace, struct thread thread,
	106	u64 tstamp, FILE *fp)
	107	{
	108	size_t printed = trace__fprintf_tstamp(trace, tstamp, fp);
	109
	110	if (trace->multiple_threads)
	111	printed += fprintf(fp, "%d ", thread->pid);
	112
	113	return printed;
	114	}
	115
	116	static int trace__process_event(struct machine machine, union perf_event event)
	117	{
	118	int ret = 0;
	119
	120	switch (event->header.type) {
	121	case PERF_RECORD_LOST:
	122	color_fprintf(stdout, PERF_COLOR_RED,
	123	"LOST %" PRIu64 " events!\n", event->lost.lost);
	124	ret = machine__process_lost_event(machine, event);
	125	default:
	126	ret = machine__process_event(machine, event);
	127	break;
	128	}
	129
	130	return ret;
	131	}
	132
	133	static int trace__tool_process(struct perf_tool *tool __maybe_unused,
	134	union perf_event *event,
	135	struct perf_sample *sample __maybe_unused,
	136	struct machine *machine)
	137	{
	138	return trace__process_event(machine, event);
	139	}
	140
	141	static int trace__symbols_init(struct trace trace, struct perf_evlist evlist)
	142	{
	143	int err = symbol__init();
	144
	145	if (err)
	146	return err;
	147
	148	machine__init(&trace->host, "", HOST_KERNEL_ID);
	149	machine__create_kernel_maps(&trace->host);
	150
	151	if (perf_target__has_task(&trace->opts.target)) {
	152	err = perf_event__synthesize_thread_map(NULL, evlist->threads,
	153	trace__tool_process,
	154	&trace->host);
	155	} else {
	156	err = perf_event__synthesize_threads(NULL, trace__tool_process,
	157	&trace->host);
	158	}
	159
	160	if (err)
	161	symbol__exit();
	162
	163	return err;
	164	}
	165
514f1c67 ACM	166	static int trace__read_syscall_info(struct trace *trace, int id)
	167	{
	168	char tp_name[128];
	169	struct syscall *sc;
3a531260 ACM	170	const char *name = audit_syscall_to_name(id, trace->audit_machine);
	171
	172	if (name == NULL)
	173	return -1;
514f1c67 ACM	174
	175	if (id > trace->syscalls.max) {
	176	struct syscall nsyscalls = realloc(trace->syscalls.table, (id + 1) sizeof(*sc));
	177
	178	if (nsyscalls == NULL)
	179	return -1;
	180
	181	if (trace->syscalls.max != -1) {
	182	memset(nsyscalls + trace->syscalls.max + 1, 0,
	183	(id - trace->syscalls.max) * sizeof(*sc));
	184	} else {
	185	memset(nsyscalls, 0, (id + 1) * sizeof(*sc));
	186	}
	187
	188	trace->syscalls.table = nsyscalls;
	189	trace->syscalls.max = id;
	190	}
	191
	192	sc = trace->syscalls.table + id;
3a531260 ACM	193	sc->name = name;
3a531260 ACM	194	sc->fmt = syscall_fmt__find(sc->name);
514f1c67	195
aec1930b	196	snprintf(tp_name, sizeof(tp_name), "sys_enter_%s", sc->name);
514f1c67	197	sc->tp_format = event_format__new("syscalls", tp_name);
aec1930b ACM	198
	199	if (sc->tp_format == NULL && sc->fmt && sc->fmt->alias) {
	200	snprintf(tp_name, sizeof(tp_name), "sys_enter_%s", sc->fmt->alias);
	201	sc->tp_format = event_format__new("syscalls", tp_name);
	202	}
514f1c67 ACM	203
	204	return sc->tp_format != NULL ? 0 : -1;
	205	}
	206
752fde44 ACM	207	static size_t syscall__scnprintf_args(struct syscall sc, char bf, size_t size,
752fde44 ACM	208	unsigned long *args)
514f1c67 ACM	209	{
	210	int i = 0;
	211	size_t printed = 0;
	212
	213	if (sc->tp_format != NULL) {
	214	struct format_field *field;
	215
	216	for (field = sc->tp_format->format.fields->next; field; field = field->next) {
752fde44 ACM	217	printed += scnprintf(bf + printed, size - printed,
	218	"%s%s: %ld", printed ? ", " : "",
	219	field->name, args[i++]);
514f1c67 ACM	220	}
	221	} else {
	222	while (i < 6) {
752fde44 ACM	223	printed += scnprintf(bf + printed, size - printed,
	224	"%sarg%d: %ld",
	225	printed ? ", " : "", i, args[i]);
514f1c67 ACM	226	++i;
	227	}
	228	}
	229
	230	return printed;
	231	}
	232
ba3d7dee ACM	233	typedef int (tracepoint_handler)(struct trace trace, struct perf_evsel *evsel,
	234	struct perf_sample *sample);
	235
	236	static struct syscall trace__syscall_info(struct trace trace,
	237	struct perf_evsel *evsel,
	238	struct perf_sample *sample)
	239	{
	240	int id = perf_evsel__intval(evsel, sample, "id");
	241
	242	if (id < 0) {
	243	printf("Invalid syscall %d id, skipping...\n", id);
	244	return NULL;
	245	}
	246
	247	if ((id > trace->syscalls.max \|\| trace->syscalls.table[id].name == NULL) &&
	248	trace__read_syscall_info(trace, id))
	249	goto out_cant_read;
	250
	251	if ((id > trace->syscalls.max \|\| trace->syscalls.table[id].name == NULL))
	252	goto out_cant_read;
	253
	254	return &trace->syscalls.table[id];
	255
	256	out_cant_read:
	257	printf("Problems reading syscall %d information\n", id);
	258	return NULL;
	259	}
	260
	261	static int trace__sys_enter(struct trace trace, struct perf_evsel evsel,
	262	struct perf_sample *sample)
	263	{
752fde44	264	char *msg;
ba3d7dee	265	void *args;
752fde44 ACM	266	size_t printed = 0;
752fde44 ACM	267	struct thread *thread = machine__findnew_thread(&trace->host, sample->tid);
ba3d7dee	268	struct syscall *sc = trace__syscall_info(trace, evsel, sample);
752fde44	269	struct thread_trace *ttrace = thread__trace(thread);
ba3d7dee	270
752fde44	271	if (ttrace == NULL \|\| sc == NULL)
ba3d7dee ACM	272	return -1;
	273
	274	args = perf_evsel__rawptr(evsel, sample, "args");
	275	if (args == NULL) {
	276	printf("Problems reading syscall arguments\n");
	277	return -1;
	278	}
	279
752fde44 ACM	280	ttrace = thread->priv;
	281
	282	if (ttrace->entry_str == NULL) {
	283	ttrace->entry_str = malloc(1024);
	284	if (!ttrace->entry_str)
	285	return -1;
	286	}
	287
	288	ttrace->entry_time = sample->time;
	289	msg = ttrace->entry_str;
	290	printed += scnprintf(msg + printed, 1024 - printed, "%s(", sc->name);
	291
	292	printed += syscall__scnprintf_args(sc, msg + printed, 1024 - printed, args);
	293
	294	if (!strcmp(sc->name, "exit_group") \|\| !strcmp(sc->name, "exit")) {
	295	trace__fprintf_entry_head(trace, thread, sample->time, stdout);
	296	printf("%-70s\n", ttrace->entry_str);
	297	} else
	298	ttrace->entry_pending = true;
ba3d7dee ACM	299
	300	return 0;
	301	}
	302
	303	static int trace__sys_exit(struct trace trace, struct perf_evsel evsel,
	304	struct perf_sample *sample)
	305	{
	306	int ret;
752fde44 ACM	307	struct thread *thread = machine__findnew_thread(&trace->host, sample->tid);
752fde44 ACM	308	struct thread_trace *ttrace = thread__trace(thread);
ba3d7dee ACM	309	struct syscall *sc = trace__syscall_info(trace, evsel, sample);
ba3d7dee ACM	310
752fde44	311	if (ttrace == NULL \|\| sc == NULL)
ba3d7dee ACM	312	return -1;
	313
	314	ret = perf_evsel__intval(evsel, sample, "ret");
	315
752fde44 ACM	316	ttrace = thread->priv;
	317
	318	ttrace->exit_time = sample->time;
	319
	320	trace__fprintf_entry_head(trace, thread, sample->time, stdout);
	321
	322	if (ttrace->entry_pending) {
	323	printf("%-70s", ttrace->entry_str);
	324	} else {
	325	printf(" ... [");
	326	color_fprintf(stdout, PERF_COLOR_YELLOW, "continued");
	327	printf("]: %s()", sc->name);
	328	}
	329
ba3d7dee ACM	330	if (ret < 0 && sc->fmt && sc->fmt->errmsg) {
	331	char bf[256];
	332	const char *emsg = strerror_r(-ret, bf, sizeof(bf)),
	333	*e = audit_errno_to_name(-ret);
	334
	335	printf(") = -1 %s %s", e, emsg);
	336	} else if (ret == 0 && sc->fmt && sc->fmt->timeout)
	337	printf(") = 0 Timeout");
	338	else
	339	printf(") = %d", ret);
	340
	341	putchar('\n');
752fde44 ACM	342
	343	ttrace->entry_pending = false;
	344
ba3d7dee ACM	345	return 0;
	346	}
	347
f15eb531	348	static int trace__run(struct trace trace, int argc, const char *argv)
514f1c67 ACM	349	{
514f1c67 ACM	350	struct perf_evlist *evlist = perf_evlist__new(NULL, NULL);
ba3d7dee	351	struct perf_evsel *evsel;
514f1c67	352	int err = -1, i, nr_events = 0, before;
f15eb531	353	const bool forks = argc > 0;
514f1c67 ACM	354
	355	if (evlist == NULL) {
	356	printf("Not enough memory to run!\n");
	357	goto out;
	358	}
	359
39876e7d ACM	360	if (perf_evlist__add_newtp(evlist, "raw_syscalls", "sys_enter", trace__sys_enter) \|\|
	361	perf_evlist__add_newtp(evlist, "raw_syscalls", "sys_exit", trace__sys_exit)) {
	362	printf("Couldn't read the raw_syscalls tracepoints information!\n");
514f1c67 ACM	363	goto out_delete_evlist;
	364	}
	365
514f1c67 ACM	366	err = perf_evlist__create_maps(evlist, &trace->opts.target);
	367	if (err < 0) {
	368	printf("Problems parsing the target to trace, check your options!\n");
	369	goto out_delete_evlist;
	370	}
	371
752fde44 ACM	372	err = trace__symbols_init(trace, evlist);
	373	if (err < 0) {
	374	printf("Problems initializing symbol libraries!\n");
	375	goto out_delete_evlist;
	376	}
	377
514f1c67 ACM	378	perf_evlist__config_attrs(evlist, &trace->opts);
514f1c67 ACM	379
f15eb531 NK	380	signal(SIGCHLD, sig_handler);
	381	signal(SIGINT, sig_handler);
	382
	383	if (forks) {
	384	err = perf_evlist__prepare_workload(evlist, &trace->opts, argv);
	385	if (err < 0) {
	386	printf("Couldn't run the workload!\n");
	387	goto out_delete_evlist;
	388	}
	389	}
	390
514f1c67 ACM	391	err = perf_evlist__open(evlist);
	392	if (err < 0) {
	393	printf("Couldn't create the events: %s\n", strerror(errno));
	394	goto out_delete_evlist;
	395	}
	396
	397	err = perf_evlist__mmap(evlist, UINT_MAX, false);
	398	if (err < 0) {
	399	printf("Couldn't mmap the events: %s\n", strerror(errno));
	400	goto out_delete_evlist;
	401	}
	402
	403	perf_evlist__enable(evlist);
f15eb531 NK	404
	405	if (forks)
	406	perf_evlist__start_workload(evlist);
	407
752fde44	408	trace->multiple_threads = evlist->threads->map[0] == -1 \|\| evlist->threads->nr > 1;
514f1c67 ACM	409	again:
	410	before = nr_events;
	411
	412	for (i = 0; i < evlist->nr_mmaps; i++) {
	413	union perf_event *event;
	414
	415	while ((event = perf_evlist__mmap_read(evlist, i)) != NULL) {
	416	const u32 type = event->header.type;
ba3d7dee	417	tracepoint_handler handler;
514f1c67	418	struct perf_sample sample;
514f1c67 ACM	419
	420	++nr_events;
	421
514f1c67 ACM	422	err = perf_evlist__parse_sample(evlist, event, &sample);
	423	if (err) {
	424	printf("Can't parse sample, err = %d, skipping...\n", err);
	425	continue;
	426	}
	427
752fde44 ACM	428	if (trace->base_time == 0)
	429	trace->base_time = sample.time;
	430
	431	if (type != PERF_RECORD_SAMPLE) {
	432	trace__process_event(&trace->host, event);
	433	continue;
	434	}
	435
514f1c67 ACM	436	evsel = perf_evlist__id2evsel(evlist, sample.id);
	437	if (evsel == NULL) {
	438	printf("Unknown tp ID %" PRIu64 ", skipping...\n", sample.id);
	439	continue;
	440	}
	441
752fde44 ACM	442	if (sample.raw_data == NULL) {
	443	printf("%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
	444	perf_evsel__name(evsel), sample.tid,
	445	sample.cpu, sample.raw_size);
	446	continue;
	447	}
514f1c67	448
fc551f8d ACM	449	if (sample.raw_data == NULL) {
	450	printf("%s sample with no payload for tid: %d, cpu %d, raw_size=%d, skipping...\n",
	451	perf_evsel__name(evsel), sample.tid,
	452	sample.cpu, sample.raw_size);
	453	continue;
	454	}
	455
ba3d7dee ACM	456	handler = evsel->handler.func;
ba3d7dee ACM	457	handler(trace, evsel, &sample);
514f1c67 ACM	458	}
	459	}
	460
f15eb531 NK	461	if (nr_events == before) {
	462	if (done)
	463	goto out_delete_evlist;
	464
514f1c67	465	poll(evlist->pollfd, evlist->nr_fds, -1);
f15eb531 NK	466	}
	467
	468	if (done)
	469	perf_evlist__disable(evlist);
514f1c67 ACM	470
	471	goto again;
	472
	473	out_delete_evlist:
	474	perf_evlist__delete(evlist);
	475	out:
	476	return err;
	477	}
	478
	479	int cmd_trace(int argc, const char *argv, const char prefix __maybe_unused)
	480	{
	481	const char * const trace_usage[] = {
f15eb531 NK	482	"perf trace [<options>] [<command>]",
f15eb531 NK	483	"perf trace [<options>] -- <command> [<options>]",
514f1c67 ACM	484	NULL
	485	};
	486	struct trace trace = {
	487	.audit_machine = audit_detect_machine(),
	488	.syscalls = {
	489	. max = -1,
	490	},
	491	.opts = {
	492	.target = {
	493	.uid = UINT_MAX,
	494	.uses_mmap = true,
	495	},
	496	.user_freq = UINT_MAX,
	497	.user_interval = ULLONG_MAX,
	498	.no_delay = true,
	499	.mmap_pages = 1024,
	500	},
	501	};
	502	const struct option trace_options[] = {
	503	OPT_STRING('p', "pid", &trace.opts.target.pid, "pid",
	504	"trace events on existing process id"),
	505	OPT_STRING(0, "tid", &trace.opts.target.tid, "tid",
	506	"trace events on existing thread id"),
	507	OPT_BOOLEAN(0, "all-cpus", &trace.opts.target.system_wide,
	508	"system-wide collection from all CPUs"),
	509	OPT_STRING(0, "cpu", &trace.opts.target.cpu_list, "cpu",
	510	"list of cpus to monitor"),
	511	OPT_BOOLEAN(0, "no-inherit", &trace.opts.no_inherit,
	512	"child tasks do not inherit counters"),
	513	OPT_UINTEGER(0, "mmap-pages", &trace.opts.mmap_pages,
	514	"number of mmap data pages"),
	515	OPT_STRING(0, "uid", &trace.opts.target.uid_str, "user",
	516	"user to profile"),
	517	OPT_END()
	518	};
	519	int err;
32caf0d1	520	char bf[BUFSIZ];
514f1c67 ACM	521
514f1c67 ACM	522	argc = parse_options(argc, argv, trace_options, trace_usage, 0);
514f1c67	523
32caf0d1 NK	524	err = perf_target__validate(&trace.opts.target);
	525	if (err) {
	526	perf_target__strerror(&trace.opts.target, err, bf, sizeof(bf));
	527	printf("%s", bf);
	528	return err;
	529	}
	530
514f1c67 ACM	531	err = perf_target__parse_uid(&trace.opts.target);
514f1c67 ACM	532	if (err) {
514f1c67 ACM	533	perf_target__strerror(&trace.opts.target, err, bf, sizeof(bf));
	534	printf("%s", bf);
	535	return err;
	536	}
	537
f15eb531	538	if (!argc && perf_target__none(&trace.opts.target))
ee76120e NK	539	trace.opts.target.system_wide = true;
ee76120e NK	540
f15eb531	541	return trace__run(&trace, argc, argv);
514f1c67	542	}