tools/perf/builtin-annotate.c

   1 /*
   2  * builtin-annotate.c
   3  *
   4  * Builtin annotate command: Analyze the perf.data input file,
   5  * look up and read DSOs and symbol information and display
   6  * a histogram of results, along various sorting keys.
   7  */
   8 #include "builtin.h"
   9
  10 #include "util/util.h"
  11
  12 #include "util/color.h"
  13 #include <linux/list.h>
  14 #include "util/cache.h"
  15 #include <linux/rbtree.h>
  16 #include "util/symbol.h"
  17 #include "util/string.h"
  18
  19 #include "perf.h"
  20 #include "util/debug.h"
  21
  22 #include "util/event.h"
  23 #include "util/parse-options.h"
  24 #include "util/parse-events.h"
  25 #include "util/thread.h"
  26 #include "util/sort.h"
  27 #include "util/hist.h"
  28 #include "util/session.h"
  29 #include "util/data_map.h"
  30
  31 static char             const *input_name = "perf.data";
  32
  33 static int              force;
  34
  35 static int              full_paths;
  36
  37 static int              print_line;
  38
  39 struct sym_hist {
  40         u64             sum;
  41         u64             ip[0];
  42 };
  43
  44 struct sym_ext {
  45         struct rb_node  node;
  46         double          percent;
  47         char            *path;
  48 };
  49
  50 struct sym_priv {
  51         struct sym_hist *hist;
  52         struct sym_ext  *ext;
  53 };
  54
  55 static struct symbol_conf symbol_conf = {
  56         .priv_size        = sizeof(struct sym_priv),
  57         .try_vmlinux_path = true,
  58 };
  59
  60 static const char *sym_hist_filter;
  61
  62 static int symbol_filter(struct map *map __used, struct symbol *sym)
  63 {
  64         if (sym_hist_filter == NULL ||
  65             strcmp(sym->name, sym_hist_filter) == 0) {
  66                 struct sym_priv *priv = symbol__priv(sym);
  67                 const int size = (sizeof(*priv->hist) +
  68                                   (sym->end - sym->start) * sizeof(u64));
  69
  70                 priv->hist = malloc(size);
  71                 if (priv->hist)
  72                         memset(priv->hist, 0, size);
  73                 return 0;
  74         }
  75         /*
  76          * FIXME: We should really filter it out, as we don't want to go thru symbols
  77          * we're not interested, and if a DSO ends up with no symbols, delete it too,
  78          * but right now the kernel loading routines in symbol.c bail out if no symbols
  79          * are found, fix it later.
  80          */
  81         return 0;
  82 }
  83
  84 /*
  85  * collect histogram counts
  86  */
  87 static void hist_hit(struct hist_entry *he, u64 ip)
  88 {
  89         unsigned int sym_size, offset;
  90         struct symbol *sym = he->sym;
  91         struct sym_priv *priv;
  92         struct sym_hist *h;
  93
  94         he->count++;
  95
  96         if (!sym || !he->map)
  97                 return;
  98
  99         priv = symbol__priv(sym);
 100         if (!priv->hist)
 101                 return;
 102
 103         sym_size = sym->end - sym->start;
 104         offset = ip - sym->start;
 105
 106         if (verbose)
 107                 fprintf(stderr, "%s: ip=%Lx\n", __func__,
 108                         he->map->unmap_ip(he->map, ip));
 109
 110         if (offset >= sym_size)
 111                 return;
 112
 113         h = priv->hist;
 114         h->sum++;
 115         h->ip[offset]++;
 116
 117         if (verbose >= 3)
 118                 printf("%p %s: count++ [ip: %p, %08Lx] => %Ld\n",
 119                         (void *)(unsigned long)he->sym->start,
 120                         he->sym->name,
 121                         (void *)(unsigned long)ip, ip - he->sym->start,
 122                         h->ip[offset]);
 123 }
 124
 125 static int hist_entry__add(struct addr_location *al, u64 count)
 126 {
 127         bool hit;
 128         struct hist_entry *he = __hist_entry__add(al, NULL, count, &hit);
 129         if (he == NULL)
 130                 return -ENOMEM;
 131         hist_hit(he, al->addr);
 132         return 0;
 133 }
 134
 135 static int process_sample_event(event_t *event, struct perf_session *session __used)
 136 {
 137         struct addr_location al;
 138
 139         dump_printf("(IP, %d): %d: %p\n", event->header.misc,
 140                     event->ip.pid, (void *)(long)event->ip.ip);
 141
 142         if (event__preprocess_sample(event, &al, symbol_filter) < 0) {
 143                 fprintf(stderr, "problem processing %d event, skipping it.\n",
 144                         event->header.type);
 145                 return -1;
 146         }
 147
 148         if (hist_entry__add(&al, 1)) {
 149                 fprintf(stderr, "problem incrementing symbol count, "
 150                                 "skipping event\n");
 151                 return -1;
 152         }
 153
 154         return 0;
 155 }
 156
 157 static int parse_line(FILE *file, struct hist_entry *he, u64 len)
 158 {
 159         struct symbol *sym = he->sym;
 160         char *line = NULL, *tmp, *tmp2;
 161         static const char *prev_line;
 162         static const char *prev_color;
 163         unsigned int offset;
 164         size_t line_len;
 165         u64 start;
 166         s64 line_ip;
 167         int ret;
 168         char *c;
 169
 170         if (getline(&line, &line_len, file) < 0)
 171                 return -1;
 172         if (!line)
 173                 return -1;
 174
 175         c = strchr(line, '\n');
 176         if (c)
 177                 *c = 0;
 178
 179         line_ip = -1;
 180         offset = 0;
 181         ret = -2;
 182
 183         /*
 184          * Strip leading spaces:
 185          */
 186         tmp = line;
 187         while (*tmp) {
 188                 if (*tmp != ' ')
 189                         break;
 190                 tmp++;
 191         }
 192
 193         if (*tmp) {
 194                 /*
 195                  * Parse hexa addresses followed by ':'
 196                  */
 197                 line_ip = strtoull(tmp, &tmp2, 16);
 198                 if (*tmp2 != ':')
 199                         line_ip = -1;
 200         }
 201
 202         start = he->map->unmap_ip(he->map, sym->start);
 203
 204         if (line_ip != -1) {
 205                 const char *path = NULL;
 206                 unsigned int hits = 0;
 207                 double percent = 0.0;
 208                 const char *color;
 209                 struct sym_priv *priv = symbol__priv(sym);
 210                 struct sym_ext *sym_ext = priv->ext;
 211                 struct sym_hist *h = priv->hist;
 212
 213                 offset = line_ip - start;
 214                 if (offset < len)
 215                         hits = h->ip[offset];
 216
 217                 if (offset < len && sym_ext) {
 218                         path = sym_ext[offset].path;
 219                         percent = sym_ext[offset].percent;
 220                 } else if (h->sum)
 221                         percent = 100.0 * hits / h->sum;
 222
 223                 color = get_percent_color(percent);
 224
 225                 /*
 226                  * Also color the filename and line if needed, with
 227                  * the same color than the percentage. Don't print it
 228                  * twice for close colored ip with the same filename:line
 229                  */
 230                 if (path) {
 231                         if (!prev_line || strcmp(prev_line, path)
 232                                        || color != prev_color) {
 233                                 color_fprintf(stdout, color, " %s", path);
 234                                 prev_line = path;
 235                                 prev_color = color;
 236                         }
 237                 }
 238
 239                 color_fprintf(stdout, color, " %7.2f", percent);
 240                 printf(" :      ");
 241                 color_fprintf(stdout, PERF_COLOR_BLUE, "%s\n", line);
 242         } else {
 243                 if (!*line)
 244                         printf("         :\n");
 245                 else
 246                         printf("         :      %s\n", line);
 247         }
 248
 249         return 0;
 250 }
 251
 252 static struct rb_root root_sym_ext;
 253
 254 static void insert_source_line(struct sym_ext *sym_ext)
 255 {
 256         struct sym_ext *iter;
 257         struct rb_node **p = &root_sym_ext.rb_node;
 258         struct rb_node *parent = NULL;
 259
 260         while (*p != NULL) {
 261                 parent = *p;
 262                 iter = rb_entry(parent, struct sym_ext, node);
 263
 264                 if (sym_ext->percent > iter->percent)
 265                         p = &(*p)->rb_left;
 266                 else
 267                         p = &(*p)->rb_right;
 268         }
 269
 270         rb_link_node(&sym_ext->node, parent, p);
 271         rb_insert_color(&sym_ext->node, &root_sym_ext);
 272 }
 273
 274 static void free_source_line(struct hist_entry *he, int len)
 275 {
 276         struct sym_priv *priv = symbol__priv(he->sym);
 277         struct sym_ext *sym_ext = priv->ext;
 278         int i;
 279
 280         if (!sym_ext)
 281                 return;
 282
 283         for (i = 0; i < len; i++)
 284                 free(sym_ext[i].path);
 285         free(sym_ext);
 286
 287         priv->ext = NULL;
 288         root_sym_ext = RB_ROOT;
 289 }
 290
 291 /* Get the filename:line for the colored entries */
 292 static void
 293 get_source_line(struct hist_entry *he, int len, const char *filename)
 294 {
 295         struct symbol *sym = he->sym;
 296         u64 start;
 297         int i;
 298         char cmd[PATH_MAX * 2];
 299         struct sym_ext *sym_ext;
 300         struct sym_priv *priv = symbol__priv(sym);
 301         struct sym_hist *h = priv->hist;
 302
 303         if (!h->sum)
 304                 return;
 305
 306         sym_ext = priv->ext = calloc(len, sizeof(struct sym_ext));
 307         if (!priv->ext)
 308                 return;
 309
 310         start = he->map->unmap_ip(he->map, sym->start);
 311
 312         for (i = 0; i < len; i++) {
 313                 char *path = NULL;
 314                 size_t line_len;
 315                 u64 offset;
 316                 FILE *fp;
 317
 318                 sym_ext[i].percent = 100.0 * h->ip[i] / h->sum;
 319                 if (sym_ext[i].percent <= 0.5)
 320                         continue;
 321
 322                 offset = start + i;
 323                 sprintf(cmd, "addr2line -e %s %016llx", filename, offset);
 324                 fp = popen(cmd, "r");
 325                 if (!fp)
 326                         continue;
 327
 328                 if (getline(&path, &line_len, fp) < 0 || !line_len)
 329                         goto next;
 330
 331                 sym_ext[i].path = malloc(sizeof(char) * line_len + 1);
 332                 if (!sym_ext[i].path)
 333                         goto next;
 334
 335                 strcpy(sym_ext[i].path, path);
 336                 insert_source_line(&sym_ext[i]);
 337
 338         next:
 339                 pclose(fp);
 340         }
 341 }
 342
 343 static void print_summary(const char *filename)
 344 {
 345         struct sym_ext *sym_ext;
 346         struct rb_node *node;
 347
 348         printf("\nSorted summary for file %s\n", filename);
 349         printf("----------------------------------------------\n\n");
 350
 351         if (RB_EMPTY_ROOT(&root_sym_ext)) {
 352                 printf(" Nothing higher than %1.1f%%\n", MIN_GREEN);
 353                 return;
 354         }
 355
 356         node = rb_first(&root_sym_ext);
 357         while (node) {
 358                 double percent;
 359                 const char *color;
 360                 char *path;
 361
 362                 sym_ext = rb_entry(node, struct sym_ext, node);
 363                 percent = sym_ext->percent;
 364                 color = get_percent_color(percent);
 365                 path = sym_ext->path;
 366
 367                 color_fprintf(stdout, color, " %7.2f %s", percent, path);
 368                 node = rb_next(node);
 369         }
 370 }
 371
 372 static void annotate_sym(struct hist_entry *he)
 373 {
 374         struct map *map = he->map;
 375         struct dso *dso = map->dso;
 376         struct symbol *sym = he->sym;
 377         const char *filename = dso->long_name, *d_filename;
 378         u64 len;
 379         char command[PATH_MAX*2];
 380         FILE *file;
 381
 382         if (!filename)
 383                 return;
 384
 385         if (verbose)
 386                 fprintf(stderr, "%s: filename=%s, sym=%s, start=%Lx, end=%Lx\n",
 387                         __func__, filename, sym->name,
 388                         map->unmap_ip(map, sym->start),
 389                         map->unmap_ip(map, sym->end));
 390
 391         if (full_paths)
 392                 d_filename = filename;
 393         else
 394                 d_filename = basename(filename);
 395
 396         len = sym->end - sym->start;
 397
 398         if (print_line) {
 399                 get_source_line(he, len, filename);
 400                 print_summary(filename);
 401         }
 402
 403         printf("\n\n------------------------------------------------\n");
 404         printf(" Percent |      Source code & Disassembly of %s\n", d_filename);
 405         printf("------------------------------------------------\n");
 406
 407         if (verbose >= 2)
 408                 printf("annotating [%p] %30s : [%p] %30s\n",
 409                        dso, dso->long_name, sym, sym->name);
 410
 411         sprintf(command, "objdump --start-address=0x%016Lx --stop-address=0x%016Lx -dS %s|grep -v %s",
 412                 map->unmap_ip(map, sym->start), map->unmap_ip(map, sym->end),
 413                 filename, filename);
 414
 415         if (verbose >= 3)
 416                 printf("doing: %s\n", command);
 417
 418         file = popen(command, "r");
 419         if (!file)
 420                 return;
 421
 422         while (!feof(file)) {
 423                 if (parse_line(file, he, len) < 0)
 424                         break;
 425         }
 426
 427         pclose(file);
 428         if (print_line)
 429                 free_source_line(he, len);
 430 }
 431
 432 static void find_annotations(void)
 433 {
 434         struct rb_node *nd;
 435
 436         for (nd = rb_first(&output_hists); nd; nd = rb_next(nd)) {
 437                 struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node);
 438                 struct sym_priv *priv;
 439
 440                 if (he->sym == NULL)
 441                         continue;
 442
 443                 priv = symbol__priv(he->sym);
 444                 if (priv->hist == NULL)
 445                         continue;
 446
 447                 annotate_sym(he);
 448                 /*
 449                  * Since we have a hist_entry per IP for the same symbol, free
 450                  * he->sym->hist to signal we already processed this symbol.
 451                  */
 452                 free(priv->hist);
 453                 priv->hist = NULL;
 454         }
 455 }
 456
 457 static struct perf_file_handler file_handler = {
 458         .process_sample_event   = process_sample_event,
 459         .process_mmap_event     = event__process_mmap,
 460         .process_comm_event     = event__process_comm,
 461         .process_fork_event     = event__process_task,
 462 };
 463
 464 static int __cmd_annotate(void)
 465 {
 466         struct perf_session *session = perf_session__new(input_name, O_RDONLY, force);
 467         struct thread *idle;
 468         int ret;
 469
 470         if (session == NULL)
 471                 return -ENOMEM;
 472
 473         idle = register_idle_thread();
 474         register_perf_file_handler(&file_handler);
 475
 476         ret = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd);
 477         if (ret)
 478                 goto out_delete;
 479
 480         if (dump_trace) {
 481                 event__print_totals();
 482                 goto out_delete;
 483         }
 484
 485         if (verbose > 3)
 486                 threads__fprintf(stdout);
 487
 488         if (verbose > 2)
 489                 dsos__fprintf(stdout);
 490
 491         collapse__resort();
 492         output__resort(event__total[0]);
 493
 494         find_annotations();
 495 out_delete:
 496         perf_session__delete(session);
 497
 498         return ret;
 499 }
 500
 501 static const char * const annotate_usage[] = {
 502         "perf annotate [<options>] <command>",
 503         NULL
 504 };
 505
 506 static const struct option options[] = {
 507         OPT_STRING('i', "input", &input_name, "file",
 508                     "input file name"),
 509         OPT_STRING('s', "symbol", &sym_hist_filter, "symbol",
 510                     "symbol to annotate"),
 511         OPT_BOOLEAN('f', "force", &force, "don't complain, do it"),
 512         OPT_BOOLEAN('v', "verbose", &verbose,
 513                     "be more verbose (show symbol address, etc)"),
 514         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
 515                     "dump raw trace in ASCII"),
 516         OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
 517                    "file", "vmlinux pathname"),
 518         OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules,
 519                     "load module symbols - WARNING: use only with -k and LIVE kernel"),
 520         OPT_BOOLEAN('l', "print-line", &print_line,
 521                     "print matching source lines (may be slow)"),
 522         OPT_BOOLEAN('P', "full-paths", &full_paths,
 523                     "Don't shorten the displayed pathnames"),
 524         OPT_END()
 525 };
 526
 527 static void setup_sorting(void)
 528 {
 529         char *tmp, *tok, *str = strdup(sort_order);
 530
 531         for (tok = strtok_r(str, ", ", &tmp);
 532                         tok; tok = strtok_r(NULL, ", ", &tmp)) {
 533                 if (sort_dimension__add(tok) < 0) {
 534                         error("Unknown --sort key: `%s'", tok);
 535                         usage_with_options(annotate_usage, options);
 536                 }
 537         }
 538
 539         free(str);
 540 }
 541
 542 int cmd_annotate(int argc, const char **argv, const char *prefix __used)
 543 {
 544         if (symbol__init(&symbol_conf) < 0)
 545                 return -1;
 546
 547         argc = parse_options(argc, argv, options, annotate_usage, 0);
 548
 549         setup_sorting();
 550
 551         if (argc) {
 552                 /*
 553                  * Special case: if there's an argument left then assume tha
 554                  * it's a symbol filter:
 555                  */
 556                 if (argc > 1)
 557                         usage_with_options(annotate_usage, options);
 558
 559                 sym_hist_filter = argv[0];
 560         }
 561
 562         setup_pager();
 563
 564         if (field_sep && *field_sep == '.') {
 565                 fputs("'.' is the only non valid --field-separator argument\n",
 566                                 stderr);
 567                 exit(129);
 568         }
 569
 570         return __cmd_annotate();
 571 }