data-convert-json.c

  1// SPDX-License-Identifier: GPL-2.0-only
  2/*
  3 * JSON export.
  4 *
  5 * Copyright (C) 2021, CodeWeavers Inc. <nfraser@codeweavers.com>
  6 */
  7
  8#include "data-convert.h"
  9
 10#include <fcntl.h>
 11#include <inttypes.h>
 12#include <sys/stat.h>
 13#include <unistd.h>
 14
 15#include "linux/compiler.h"
 16#include "linux/err.h"
 17#include "util/auxtrace.h"
 18#include "util/debug.h"
 19#include "util/dso.h"
 20#include "util/event.h"
 21#include "util/evsel.h"
 22#include "util/evlist.h"
 23#include "util/header.h"
 24#include "util/map.h"
 25#include "util/session.h"
 26#include "util/symbol.h"
 27#include "util/thread.h"
 28#include "util/tool.h"
 29
 30#ifdef HAVE_LIBTRACEEVENT
 31#include <event-parse.h>
 32#endif
 33
 34struct convert_json {
 35	struct perf_tool tool;
 36	FILE *out;
 37	bool first;
 38	u64 events_count;
 39};
 40
 41// Outputs a JSON-encoded string surrounded by quotes with characters escaped.
 42static void output_json_string(FILE *out, const char *s)
 43{
 44	fputc('"', out);
 45	while (*s) {
 46		switch (*s) {
 47
 48		// required escapes with special forms as per RFC 8259
 49		case '"':  fputs("\\\"", out); break;
 50		case '\\': fputs("\\\\", out); break;
 51		case '\b': fputs("\\b", out);  break;
 52		case '\f': fputs("\\f", out);  break;
 53		case '\n': fputs("\\n", out);  break;
 54		case '\r': fputs("\\r", out);  break;
 55		case '\t': fputs("\\t", out);  break;
 56
 57		default:
 58			// all other control characters must be escaped by hex code
 59			if (*s <= 0x1f)
 60				fprintf(out, "\\u%04x", *s);
 61			else
 62				fputc(*s, out);
 63			break;
 64		}
 65
 66		++s;
 67	}
 68	fputc('"', out);
 69}
 70
 71// Outputs an optional comma, newline and indentation to delimit a new value
 72// from the previous one in a JSON object or array.
 73static void output_json_delimiters(FILE *out, bool comma, int depth)
 74{
 75	int i;
 76
 77	if (comma)
 78		fputc(',', out);
 79	fputc('\n', out);
 80	for (i = 0; i < depth; ++i)
 81		fputc('\t', out);
 82}
 83
 84// Outputs a printf format string (with delimiter) as a JSON value.
 85__printf(4, 5)
 86static void output_json_format(FILE *out, bool comma, int depth, const char *format, ...)
 87{
 88	va_list args;
 89
 90	output_json_delimiters(out, comma, depth);
 91	va_start(args, format);
 92	vfprintf(out,  format, args);
 93	va_end(args);
 94}
 95
 96// Outputs a JSON key-value pair where the value is a string.
 97static void output_json_key_string(FILE *out, bool comma, int depth,
 98		const char *key, const char *value)
 99{
100	output_json_delimiters(out, comma, depth);
101	output_json_string(out, key);
102	fputs(": ", out);
103	output_json_string(out, value);
104}
105
106// Outputs a JSON key-value pair where the value is a printf format string.
107__printf(5, 6)
108static void output_json_key_format(FILE *out, bool comma, int depth,
109		const char *key, const char *format, ...)
110{
111	va_list args;
112
113	output_json_delimiters(out, comma, depth);
114	output_json_string(out, key);
115	fputs(": ", out);
116	va_start(args, format);
117	vfprintf(out,  format, args);
118	va_end(args);
119}
120
121static void output_sample_callchain_entry(const struct perf_tool *tool,
122		u64 ip, struct addr_location *al)
123{
124	struct convert_json *c = container_of(tool, struct convert_json, tool);
125	FILE *out = c->out;
126
127	output_json_format(out, false, 4, "{");
128	output_json_key_format(out, false, 5, "ip", "\"0x%" PRIx64 "\"", ip);
129
130	if (al && al->sym && al->sym->namelen) {
131		struct dso *dso = al->map ? map__dso(al->map) : NULL;
132
133		fputc(',', out);
134		output_json_key_string(out, false, 5, "symbol", al->sym->name);
135
136		if (dso) {
137			const char *dso_name = dso__short_name(dso);
138
139			if (dso_name && strlen(dso_name) > 0) {
140				fputc(',', out);
141				output_json_key_string(out, false, 5, "dso", dso_name);
142			}
143		}
144	}
145
146	output_json_format(out, false, 4, "}");
147}
148
149static int process_sample_event(const struct perf_tool *tool,
150				union perf_event *event __maybe_unused,
151				struct perf_sample *sample,
152				struct evsel *evsel __maybe_unused,
153				struct machine *machine)
154{
155	struct convert_json *c = container_of(tool, struct convert_json, tool);
156	FILE *out = c->out;
157	struct addr_location al;
158	u64 sample_type = __evlist__combined_sample_type(evsel->evlist);
159	u8 cpumode = PERF_RECORD_MISC_USER;
160
161	addr_location__init(&al);
162	if (machine__resolve(machine, &al, sample) < 0) {
163		pr_err("Sample resolution failed!\n");
164		addr_location__exit(&al);
165		return -1;
166	}
167
168	++c->events_count;
169
170	if (c->first)
171		c->first = false;
172	else
173		fputc(',', out);
174	output_json_format(out, false, 2, "{");
175
176	output_json_key_format(out, false, 3, "timestamp", "%" PRIi64, sample->time);
177	output_json_key_format(out, true, 3, "pid", "%i", thread__pid(al.thread));
178	output_json_key_format(out, true, 3, "tid", "%i", thread__tid(al.thread));
179
180	if ((sample_type & PERF_SAMPLE_CPU))
181		output_json_key_format(out, true, 3, "cpu", "%i", sample->cpu);
182	else if (thread__cpu(al.thread) >= 0)
183		output_json_key_format(out, true, 3, "cpu", "%i", thread__cpu(al.thread));
184
185	output_json_key_string(out, true, 3, "comm", thread__comm_str(al.thread));
186
187	output_json_key_format(out, true, 3, "callchain", "[");
188	if (sample->callchain) {
189		unsigned int i;
190		bool ok;
191		bool first_callchain = true;
192
193		for (i = 0; i < sample->callchain->nr; ++i) {
194			u64 ip = sample->callchain->ips[i];
195			struct addr_location tal;
196
197			if (ip >= PERF_CONTEXT_MAX) {
198				switch (ip) {
199				case PERF_CONTEXT_HV:
200					cpumode = PERF_RECORD_MISC_HYPERVISOR;
201					break;
202				case PERF_CONTEXT_KERNEL:
203					cpumode = PERF_RECORD_MISC_KERNEL;
204					break;
205				case PERF_CONTEXT_USER:
206					cpumode = PERF_RECORD_MISC_USER;
207					break;
208				default:
209					pr_debug("invalid callchain context: %"
210							PRId64 "\n", (s64) ip);
211					break;
212				}
213				continue;
214			}
215
216			if (first_callchain)
217				first_callchain = false;
218			else
219				fputc(',', out);
220
221			addr_location__init(&tal);
222			ok = thread__find_symbol(al.thread, cpumode, ip, &tal);
223			output_sample_callchain_entry(tool, ip, ok ? &tal : NULL);
224			addr_location__exit(&tal);
225		}
226	} else {
227		output_sample_callchain_entry(tool, sample->ip, &al);
228	}
229	output_json_format(out, false, 3, "]");
230
231#ifdef HAVE_LIBTRACEEVENT
232	if (sample->raw_data) {
233		struct tep_event *tp_format = evsel__tp_format(evsel);
234		struct tep_format_field **fields = tp_format ? tep_event_fields(tp_format) : NULL;
235
236		if (fields) {
237			int i = 0;
238
239			while (fields[i]) {
240				struct trace_seq s;
241
242				trace_seq_init(&s);
243				tep_print_field(&s, sample->raw_data, fields[i]);
244				output_json_key_string(out, true, 3, fields[i]->name, s.buffer);
245
246				i++;
247			}
248			free(fields);
249		}
250	}
251#endif
252	output_json_format(out, false, 2, "}");
253	addr_location__exit(&al);
254	return 0;
255}
256
257static void output_headers(struct perf_session *session, struct convert_json *c)
258{
259	struct stat st;
260	const struct perf_header *header = &session->header;
261	const struct perf_env *env = perf_session__env(session);
262	int ret;
263	int fd = perf_data__fd(session->data);
264	int i;
265	FILE *out = c->out;
266
267	output_json_key_format(out, false, 2, "header-version", "%u", header->version);
268
269	ret = fstat(fd, &st);
270	if (ret >= 0) {
271		time_t stctime = st.st_mtime;
272		char buf[256];
273
274		strftime(buf, sizeof(buf), "%FT%TZ", gmtime(&stctime));
275		output_json_key_string(out, true, 2, "captured-on", buf);
276	} else {
277		pr_debug("Failed to get mtime of source file, not writing captured-on");
278	}
279
280	output_json_key_format(out, true, 2, "data-offset", "%" PRIu64, header->data_offset);
281	output_json_key_format(out, true, 2, "data-size", "%" PRIu64, header->data_size);
282	output_json_key_format(out, true, 2, "feat-offset", "%" PRIu64, header->feat_offset);
283
284	output_json_key_string(out, true, 2, "hostname", env->hostname);
285	output_json_key_string(out, true, 2, "os-release", env->os_release);
286	output_json_key_string(out, true, 2, "arch", env->arch);
287
288	if (env->cpu_desc)
289		output_json_key_string(out, true, 2, "cpu-desc", env->cpu_desc);
290
291	output_json_key_string(out, true, 2, "cpuid", env->cpuid);
292	output_json_key_format(out, true, 2, "nrcpus-online", "%u", env->nr_cpus_online);
293	output_json_key_format(out, true, 2, "nrcpus-avail", "%u", env->nr_cpus_avail);
294
295	if (env->clock.enabled) {
296		output_json_key_format(out, true, 2, "clockid",
297				"%u", env->clock.clockid);
298		output_json_key_format(out, true, 2, "clock-time",
299				"%" PRIu64, env->clock.clockid_ns);
300		output_json_key_format(out, true, 2, "real-time",
301				"%" PRIu64, env->clock.tod_ns);
302	}
303
304	output_json_key_string(out, true, 2, "perf-version", env->version);
305
306	output_json_key_format(out, true, 2, "cmdline", "[");
307	for (i = 0; i < env->nr_cmdline; i++) {
308		output_json_delimiters(out, i != 0, 3);
309		output_json_string(c->out, env->cmdline_argv[i]);
310	}
311	output_json_format(out, false, 2, "]");
312}
313
314int bt_convert__perf2json(const char *input_name, const char *output_name,
315		struct perf_data_convert_opts *opts __maybe_unused)
316{
317	struct perf_session *session;
318	int fd;
319	int ret = -1;
320	struct convert_json c = {
321		.first = true,
322		.events_count = 0,
323	};
324	struct perf_data data = {
325		.mode = PERF_DATA_MODE_READ,
326		.path = input_name,
327		.force = opts->force,
328	};
329
330	perf_tool__init(&c.tool, /*ordered_events=*/true);
331	c.tool.sample         = process_sample_event;
332	c.tool.mmap           = perf_event__process_mmap;
333	c.tool.mmap2          = perf_event__process_mmap2;
334	c.tool.comm           = perf_event__process_comm;
335	c.tool.namespaces     = perf_event__process_namespaces;
336	c.tool.cgroup         = perf_event__process_cgroup;
337	c.tool.exit           = perf_event__process_exit;
338	c.tool.fork           = perf_event__process_fork;
339	c.tool.lost           = perf_event__process_lost;
340#ifdef HAVE_LIBTRACEEVENT
341	c.tool.tracing_data   = perf_event__process_tracing_data;
342#endif
343	c.tool.build_id       = perf_event__process_build_id;
344	c.tool.id_index       = perf_event__process_id_index;
345	c.tool.auxtrace_info  = perf_event__process_auxtrace_info;
346	c.tool.auxtrace       = perf_event__process_auxtrace;
347	c.tool.event_update   = perf_event__process_event_update;
348	c.tool.ordering_requires_timestamps = true;
349
350	if (opts->all) {
351		pr_err("--all is currently unsupported for JSON output.\n");
352		goto err;
353	}
354	if (opts->tod) {
355		pr_err("--tod is currently unsupported for JSON output.\n");
356		goto err;
357	}
358
359	fd = open(output_name, O_CREAT | O_WRONLY | (opts->force ? O_TRUNC : O_EXCL), 0666);
360	if (fd == -1) {
361		if (errno == EEXIST)
362			pr_err("Output file exists. Use --force to overwrite it.\n");
363		else
364			pr_err("Error opening output file!\n");
365		goto err;
366	}
367
368	c.out = fdopen(fd, "w");
369	if (!c.out) {
370		fprintf(stderr, "Error opening output file!\n");
371		close(fd);
372		goto err;
373	}
374
375	session = perf_session__new(&data, &c.tool);
376	if (IS_ERR(session)) {
377		fprintf(stderr, "Error creating perf session!\n");
378		goto err_fclose;
379	}
380	if (symbol__init(perf_session__env(session)) < 0) {
381		fprintf(stderr, "Symbol init error!\n");
382		goto err_session_delete;
383	}
384
385	// The opening brace is printed manually because it isn't delimited from a
386	// previous value (i.e. we don't want a leading newline)
387	fputc('{', c.out);
388
389	// Version number for future-proofing. Most additions should be able to be
390	// done in a backwards-compatible way so this should only need to be bumped
391	// if some major breaking change must be made.
392	output_json_format(c.out, false, 1, "\"linux-perf-json-version\": 1");
393
394	// Output headers
395	output_json_format(c.out, true, 1, "\"headers\": {");
396	output_headers(session, &c);
397	output_json_format(c.out, false, 1, "}");
398
399	// Output samples
400	output_json_format(c.out, true, 1, "\"samples\": [");
401	perf_session__process_events(session);
402	output_json_format(c.out, false, 1, "]");
403	output_json_format(c.out, false, 0, "}");
404	fputc('\n', c.out);
405
406	fprintf(stderr,
407			"[ perf data convert: Converted '%s' into JSON data '%s' ]\n",
408			data.path, output_name);
409
410	fprintf(stderr,
411			"[ perf data convert: Converted and wrote %.3f MB (%" PRIu64 " samples) ]\n",
412			(ftell(c.out)) / 1024.0 / 1024.0, c.events_count);
413
414	ret = 0;
415err_session_delete:
416	perf_session__delete(session);
417err_fclose:
418	fclose(c.out);
419err:
420	return ret;
421}