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/env.h" 20 #include "util/dso.h" 21 #include "util/event.h" 22 #include "util/evsel.h" 23 #include "util/evlist.h" 24 #include "util/header.h" 25 #include "util/map.h" 26 #include "util/session.h" 27 #include "util/symbol.h" 28 #include "util/thread.h" 29 #include "util/time-utils.h" 30 #include "util/tool.h" 31 32 #ifdef HAVE_LIBTRACEEVENT 33 #include <event-parse.h> 34 #endif 35 36 struct convert_json { 37 struct perf_tool tool; 38 FILE *out; 39 bool first; 40 struct perf_time_interval *ptime_range; 41 int range_size; 42 int range_num; 43 44 u64 events_count; 45 u64 skipped; 46 }; 47 48 // Outputs a JSON-encoded string surrounded by quotes with characters escaped. 49 static void output_json_string(FILE *out, const char *s) 50 { 51 fputc('"', out); 52 if (!s) 53 goto out; 54 55 while (*s) { 56 switch (*s) { 57 58 // required escapes with special forms as per RFC 8259 59 case '"': fputs("\\\"", out); break; 60 case '\\': fputs("\\\\", out); break; 61 case '\b': fputs("\\b", out); break; 62 case '\f': fputs("\\f", out); break; 63 case '\n': fputs("\\n", out); break; 64 case '\r': fputs("\\r", out); break; 65 case '\t': fputs("\\t", out); break; 66 67 default: 68 // all other control characters must be escaped by hex code 69 if (*s <= 0x1f) 70 fprintf(out, "\\u%04x", *s); 71 else 72 fputc(*s, out); 73 break; 74 } 75 76 ++s; 77 } 78 out: 79 fputc('"', out); 80 } 81 82 // Outputs an optional comma, newline and indentation to delimit a new value 83 // from the previous one in a JSON object or array. 84 static void output_json_delimiters(FILE *out, bool comma, int depth) 85 { 86 int i; 87 88 if (comma) 89 fputc(',', out); 90 fputc('\n', out); 91 for (i = 0; i < depth; ++i) 92 fputc('\t', out); 93 } 94 95 // Outputs a printf format string (with delimiter) as a JSON value. 96 __printf(4, 5) 97 static void output_json_format(FILE *out, bool comma, int depth, const char *format, ...) 98 { 99 va_list args; 100 101 output_json_delimiters(out, comma, depth); 102 va_start(args, format); 103 vfprintf(out, format, args); 104 va_end(args); 105 } 106 107 // Outputs a JSON key-value pair where the value is a string. 108 static void output_json_key_string(FILE *out, bool comma, int depth, 109 const char *key, const char *value) 110 { 111 output_json_delimiters(out, comma, depth); 112 output_json_string(out, key); 113 fputs(": ", out); 114 output_json_string(out, value); 115 } 116 117 // Outputs a JSON key-value pair where the value is a printf format string. 118 __printf(5, 6) 119 static void output_json_key_format(FILE *out, bool comma, int depth, 120 const char *key, const char *format, ...) 121 { 122 va_list args; 123 124 output_json_delimiters(out, comma, depth); 125 output_json_string(out, key); 126 fputs(": ", out); 127 va_start(args, format); 128 vfprintf(out, format, args); 129 va_end(args); 130 } 131 132 static void output_sample_callchain_entry(const struct perf_tool *tool, 133 u64 ip, struct addr_location *al) 134 { 135 struct convert_json *c = container_of(tool, struct convert_json, tool); 136 FILE *out = c->out; 137 138 output_json_format(out, false, 4, "{"); 139 output_json_key_format(out, false, 5, "ip", "\"0x%" PRIx64 "\"", ip); 140 141 if (al && al->sym && al->sym->namelen) { 142 struct dso *dso = al->map ? map__dso(al->map) : NULL; 143 144 fputc(',', out); 145 output_json_key_string(out, false, 5, "symbol", al->sym->name); 146 147 if (dso) { 148 const char *dso_name = dso__short_name(dso); 149 150 if (dso_name && strlen(dso_name) > 0) { 151 fputc(',', out); 152 output_json_key_string(out, false, 5, "dso", dso_name); 153 } 154 } 155 } 156 157 output_json_format(out, false, 4, "}"); 158 } 159 160 static int process_sample_event(const struct perf_tool *tool, 161 union perf_event *event __maybe_unused, 162 struct perf_sample *sample, 163 struct machine *machine) 164 { 165 struct convert_json *c = container_of(tool, struct convert_json, tool); 166 FILE *out = c->out; 167 struct addr_location al; 168 u64 sample_type = __evlist__combined_sample_type(sample->evsel->evlist); 169 u8 cpumode = PERF_RECORD_MISC_USER; 170 171 addr_location__init(&al); 172 if (machine__resolve(machine, &al, sample) < 0) { 173 pr_err("Sample resolution failed!\n"); 174 addr_location__exit(&al); 175 return -1; 176 } 177 178 if (perf_time__ranges_skip_sample(c->ptime_range, c->range_num, sample->time)) { 179 ++c->skipped; 180 addr_location__exit(&al); 181 return 0; 182 } 183 184 ++c->events_count; 185 186 if (c->first) 187 c->first = false; 188 else 189 fputc(',', out); 190 output_json_format(out, false, 2, "{"); 191 192 output_json_key_format(out, false, 3, "timestamp", "%" PRIi64, sample->time); 193 output_json_key_format(out, true, 3, "pid", "%i", thread__pid(al.thread)); 194 output_json_key_format(out, true, 3, "tid", "%i", thread__tid(al.thread)); 195 196 if ((sample_type & PERF_SAMPLE_CPU)) 197 output_json_key_format(out, true, 3, "cpu", "%i", sample->cpu); 198 else if (thread__cpu(al.thread) >= 0) 199 output_json_key_format(out, true, 3, "cpu", "%i", thread__cpu(al.thread)); 200 201 output_json_key_string(out, true, 3, "comm", thread__comm_str(al.thread)); 202 203 output_json_key_format(out, true, 3, "callchain", "["); 204 if (sample->callchain) { 205 unsigned int i; 206 bool ok; 207 bool first_callchain = true; 208 209 for (i = 0; i < sample->callchain->nr; ++i) { 210 u64 ip = sample->callchain->ips[i]; 211 struct addr_location tal; 212 213 if (ip >= PERF_CONTEXT_MAX) { 214 switch (ip) { 215 case PERF_CONTEXT_HV: 216 cpumode = PERF_RECORD_MISC_HYPERVISOR; 217 break; 218 case PERF_CONTEXT_KERNEL: 219 cpumode = PERF_RECORD_MISC_KERNEL; 220 break; 221 case PERF_CONTEXT_USER: 222 cpumode = PERF_RECORD_MISC_USER; 223 break; 224 default: 225 pr_debug("invalid callchain context: %" 226 PRId64 "\n", (s64) ip); 227 break; 228 } 229 continue; 230 } 231 232 if (first_callchain) 233 first_callchain = false; 234 else 235 fputc(',', out); 236 237 addr_location__init(&tal); 238 ok = thread__find_symbol(al.thread, cpumode, ip, &tal); 239 output_sample_callchain_entry(tool, ip, ok ? &tal : NULL); 240 addr_location__exit(&tal); 241 } 242 } else { 243 output_sample_callchain_entry(tool, sample->ip, &al); 244 } 245 output_json_format(out, false, 3, "]"); 246 247 #ifdef HAVE_LIBTRACEEVENT 248 if (sample->raw_data) { 249 struct tep_event *tp_format = evsel__tp_format(sample->evsel); 250 struct tep_format_field **fields = tp_format ? tep_event_fields(tp_format) : NULL; 251 252 if (fields) { 253 int i = 0; 254 255 while (fields[i]) { 256 struct trace_seq s; 257 258 trace_seq_init(&s); 259 tep_print_field(&s, sample->raw_data, fields[i]); 260 output_json_key_string(out, true, 3, fields[i]->name, s.buffer); 261 262 i++; 263 } 264 free(fields); 265 } 266 } 267 #endif 268 output_json_format(out, false, 2, "}"); 269 addr_location__exit(&al); 270 return 0; 271 } 272 273 static void output_headers(struct perf_session *session, struct convert_json *c) 274 { 275 struct stat st; 276 const struct perf_header *header = &session->header; 277 struct perf_env *env = perf_session__env(session); 278 int ret; 279 int fd = perf_data__fd(session->data); 280 int i; 281 FILE *out = c->out; 282 283 output_json_key_format(out, false, 2, "header-version", "%u", header->version); 284 285 ret = fstat(fd, &st); 286 if (ret >= 0) { 287 time_t stctime = st.st_mtime; 288 char buf[256]; 289 290 strftime(buf, sizeof(buf), "%FT%TZ", gmtime(&stctime)); 291 output_json_key_string(out, true, 2, "captured-on", buf); 292 } else { 293 pr_debug("Failed to get mtime of source file, not writing captured-on"); 294 } 295 296 output_json_key_format(out, true, 2, "data-offset", "%" PRIu64, header->data_offset); 297 output_json_key_format(out, true, 2, "data-size", "%" PRIu64, header->data_size); 298 output_json_key_format(out, true, 2, "feat-offset", "%" PRIu64, header->feat_offset); 299 300 output_json_key_string(out, true, 2, "hostname", env->hostname); 301 output_json_key_string(out, true, 2, "os-release", 302 perf_env__os_release(env)); 303 output_json_key_string(out, true, 2, "arch", env->arch); 304 305 if (env->cpu_desc) 306 output_json_key_string(out, true, 2, "cpu-desc", env->cpu_desc); 307 308 output_json_key_string(out, true, 2, "cpuid", env->cpuid); 309 output_json_key_format(out, true, 2, "nrcpus-online", "%u", env->nr_cpus_online); 310 output_json_key_format(out, true, 2, "nrcpus-avail", "%u", env->nr_cpus_avail); 311 312 if (env->clock.enabled) { 313 output_json_key_format(out, true, 2, "clockid", 314 "%u", env->clock.clockid); 315 output_json_key_format(out, true, 2, "clock-time", 316 "%" PRIu64, env->clock.clockid_ns); 317 output_json_key_format(out, true, 2, "real-time", 318 "%" PRIu64, env->clock.tod_ns); 319 } 320 321 output_json_key_string(out, true, 2, "perf-version", env->version); 322 323 output_json_key_format(out, true, 2, "cmdline", "["); 324 for (i = 0; i < env->nr_cmdline; i++) { 325 output_json_delimiters(out, i != 0, 3); 326 output_json_string(c->out, env->cmdline_argv[i]); 327 } 328 output_json_format(out, false, 2, "]"); 329 } 330 331 int bt_convert__perf2json(const char *_input_name, const char *output_name, 332 struct perf_data_convert_opts *opts __maybe_unused) 333 { 334 struct perf_session *session; 335 int fd; 336 int ret = -1; 337 struct convert_json c = { 338 .first = true, 339 .events_count = 0, 340 .ptime_range = NULL, 341 .range_size = 0, 342 .range_num = 0, 343 .skipped = 0, 344 }; 345 struct perf_data data = { 346 .mode = PERF_DATA_MODE_READ, 347 .path = _input_name, 348 .force = opts->force, 349 }; 350 351 perf_tool__init(&c.tool, /*ordered_events=*/true); 352 c.tool.sample = process_sample_event; 353 c.tool.mmap = perf_event__process_mmap; 354 c.tool.mmap2 = perf_event__process_mmap2; 355 c.tool.comm = perf_event__process_comm; 356 c.tool.namespaces = perf_event__process_namespaces; 357 c.tool.cgroup = perf_event__process_cgroup; 358 c.tool.exit = perf_event__process_exit; 359 c.tool.fork = perf_event__process_fork; 360 c.tool.lost = perf_event__process_lost; 361 #ifdef HAVE_LIBTRACEEVENT 362 c.tool.tracing_data = perf_event__process_tracing_data; 363 #endif 364 c.tool.build_id = perf_event__process_build_id; 365 c.tool.id_index = perf_event__process_id_index; 366 c.tool.auxtrace_info = perf_event__process_auxtrace_info; 367 c.tool.auxtrace = perf_event__process_auxtrace; 368 c.tool.event_update = perf_event__process_event_update; 369 c.tool.attr = perf_event__process_attr; 370 c.tool.feature = perf_event__process_feature; 371 c.tool.ordering_requires_timestamps = true; 372 373 if (opts->all) { 374 pr_err("--all is currently unsupported for JSON output.\n"); 375 goto err; 376 } 377 if (opts->tod) { 378 pr_err("--tod is currently unsupported for JSON output.\n"); 379 goto err; 380 } 381 382 fd = open(output_name, O_CREAT | O_WRONLY | (opts->force ? O_TRUNC : O_EXCL), 0666); 383 if (fd == -1) { 384 if (errno == EEXIST) 385 pr_err("Output file exists. Use --force to overwrite it.\n"); 386 else 387 pr_err("Error opening output file!\n"); 388 goto err; 389 } 390 391 c.out = fdopen(fd, "w"); 392 if (!c.out) { 393 fprintf(stderr, "Error opening output file!\n"); 394 close(fd); 395 goto err; 396 } 397 398 session = perf_session__new(&data, &c.tool); 399 if (IS_ERR(session)) { 400 fprintf(stderr, "Error creating perf session!\n"); 401 goto err_fclose; 402 } 403 if (symbol__init(perf_session__env(session)) < 0) { 404 fprintf(stderr, "Symbol init error!\n"); 405 goto err_session_delete; 406 } 407 408 if (opts->time_str) { 409 ret = perf_time__parse_for_ranges(opts->time_str, session, 410 &c.ptime_range, 411 &c.range_size, 412 &c.range_num); 413 if (ret < 0) 414 goto err_session_delete; 415 } 416 417 // The opening brace is printed manually because it isn't delimited from a 418 // previous value (i.e. we don't want a leading newline) 419 fputc('{', c.out); 420 421 // Version number for future-proofing. Most additions should be able to be 422 // done in a backwards-compatible way so this should only need to be bumped 423 // if some major breaking change must be made. 424 output_json_format(c.out, false, 1, "\"linux-perf-json-version\": 1"); 425 426 // Output headers 427 output_json_format(c.out, true, 1, "\"headers\": {"); 428 output_headers(session, &c); 429 output_json_format(c.out, false, 1, "}"); 430 431 // Output samples 432 output_json_format(c.out, true, 1, "\"samples\": ["); 433 perf_session__process_events(session); 434 output_json_format(c.out, false, 1, "]"); 435 output_json_format(c.out, false, 0, "}"); 436 fputc('\n', c.out); 437 438 fprintf(stderr, "[ perf data convert: Converted '%s' into JSON data '%s' ]\n", 439 data.path, output_name); 440 441 fprintf(stderr, 442 "[ perf data convert: Converted and wrote %.3f MB (%" PRIu64 " samples) ]\n", 443 (ftell(c.out)) / 1024.0 / 1024.0, c.events_count); 444 445 if (c.skipped) { 446 fprintf(stderr, "[ perf data convert: Skipped %" PRIu64 " samples ]\n", 447 c.skipped); 448 } 449 450 ret = 0; 451 452 if (c.ptime_range) 453 zfree(&c.ptime_range); 454 455 err_session_delete: 456 perf_session__delete(session); 457 err_fclose: 458 fclose(c.out); 459 err: 460 return ret; 461 } 462