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