tools/perf/builtin-lock.c

   1 #include "builtin.h"
   2 #include "perf.h"
   3
   4 #include "util/util.h"
   5 #include "util/cache.h"
   6 #include "util/symbol.h"
   7 #include "util/thread.h"
   8 #include "util/header.h"
   9
  10 #include "util/parse-options.h"
  11 #include "util/trace-event.h"
  12
  13 #include "util/debug.h"
  14 #include "util/session.h"
  15
  16 #include <sys/types.h>
  17 #include <sys/prctl.h>
  18 #include <semaphore.h>
  19 #include <pthread.h>
  20 #include <math.h>
  21 #include <limits.h>
  22
  23 #include <linux/list.h>
  24 #include <linux/hash.h>
  25
  26 static struct perf_session *session;
  27
  28 /* based on kernel/lockdep.c */
  29 #define LOCKHASH_BITS           12
  30 #define LOCKHASH_SIZE           (1UL << LOCKHASH_BITS)
  31
  32 static struct list_head lockhash_table[LOCKHASH_SIZE];
  33
  34 #define __lockhashfn(key)       hash_long((unsigned long)key, LOCKHASH_BITS)
  35 #define lockhashentry(key)      (lockhash_table + __lockhashfn((key)))
  36
  37 struct lock_stat {
  38         struct list_head        hash_entry;
  39         struct rb_node          rb;             /* used for sorting */
  40
  41         /*
  42          * FIXME: raw_field_value() returns unsigned long long,
  43          * so address of lockdep_map should be dealed as 64bit.
  44          * Is there more better solution?
  45          */
  46         void                    *addr;          /* address of lockdep_map, used as ID */
  47         char                    *name;          /* for strcpy(), we cannot use const */
  48
  49         unsigned int            nr_acquire;
  50         unsigned int            nr_acquired;
  51         unsigned int            nr_contended;
  52         unsigned int            nr_release;
  53
  54         unsigned int            nr_readlock;
  55         unsigned int            nr_trylock;
  56         /* these times are in nano sec. */
  57         u64                     wait_time_total;
  58         u64                     wait_time_min;
  59         u64                     wait_time_max;
  60
  61         int                     discard; /* flag of blacklist */
  62 };
  63
  64 /*
  65  * States of lock_seq_stat
  66  *
  67  * UNINITIALIZED is required for detecting first event of acquire.
  68  * As the nature of lock events, there is no guarantee
  69  * that the first event for the locks are acquire,
  70  * it can be acquired, contended or release.
  71  */
  72 #define SEQ_STATE_UNINITIALIZED      0         /* initial state */
  73 #define SEQ_STATE_RELEASED      1
  74 #define SEQ_STATE_ACQUIRING     2
  75 #define SEQ_STATE_ACQUIRED      3
  76 #define SEQ_STATE_READ_ACQUIRED 4
  77 #define SEQ_STATE_CONTENDED     5
  78
  79 /*
  80  * MAX_LOCK_DEPTH
  81  * Imported from include/linux/sched.h.
  82  * Should this be synchronized?
  83  */
  84 #define MAX_LOCK_DEPTH 48
  85
  86 /*
  87  * struct lock_seq_stat:
  88  * Place to put on state of one lock sequence
  89  * 1) acquire -> acquired -> release
  90  * 2) acquire -> contended -> acquired -> release
  91  * 3) acquire (with read or try) -> release
  92  * 4) Are there other patterns?
  93  */
  94 struct lock_seq_stat {
  95         struct list_head        list;
  96         int                     state;
  97         u64                     prev_event_time;
  98         void                    *addr;
  99
 100         int                     read_count;
 101 };
 102
 103 struct thread_stat {
 104         struct rb_node          rb;
 105
 106         u32                     tid;
 107         struct list_head        seq_list;
 108 };
 109
 110 static struct rb_root           thread_stats;
 111
 112 static struct thread_stat *thread_stat_find(u32 tid)
 113 {
 114         struct rb_node *node;
 115         struct thread_stat *st;
 116
 117         node = thread_stats.rb_node;
 118         while (node) {
 119                 st = container_of(node, struct thread_stat, rb);
 120                 if (st->tid == tid)
 121                         return st;
 122                 else if (tid < st->tid)
 123                         node = node->rb_left;
 124                 else
 125                         node = node->rb_right;
 126         }
 127
 128         return NULL;
 129 }
 130
 131 static void thread_stat_insert(struct thread_stat *new)
 132 {
 133         struct rb_node **rb = &thread_stats.rb_node;
 134         struct rb_node *parent = NULL;
 135         struct thread_stat *p;
 136
 137         while (*rb) {
 138                 p = container_of(*rb, struct thread_stat, rb);
 139                 parent = *rb;
 140
 141                 if (new->tid < p->tid)
 142                         rb = &(*rb)->rb_left;
 143                 else if (new->tid > p->tid)
 144                         rb = &(*rb)->rb_right;
 145                 else
 146                         BUG_ON("inserting invalid thread_stat\n");
 147         }
 148
 149         rb_link_node(&new->rb, parent, rb);
 150         rb_insert_color(&new->rb, &thread_stats);
 151 }
 152
 153 static struct thread_stat *thread_stat_findnew_after_first(u32 tid)
 154 {
 155         struct thread_stat *st;
 156
 157         st = thread_stat_find(tid);
 158         if (st)
 159                 return st;
 160
 161         st = zalloc(sizeof(struct thread_stat));
 162         if (!st)
 163                 die("memory allocation failed\n");
 164
 165         st->tid = tid;
 166         INIT_LIST_HEAD(&st->seq_list);
 167
 168         thread_stat_insert(st);
 169
 170         return st;
 171 }
 172
 173 static struct thread_stat *thread_stat_findnew_first(u32 tid);
 174 static struct thread_stat *(*thread_stat_findnew)(u32 tid) =
 175         thread_stat_findnew_first;
 176
 177 static struct thread_stat *thread_stat_findnew_first(u32 tid)
 178 {
 179         struct thread_stat *st;
 180
 181         st = zalloc(sizeof(struct thread_stat));
 182         if (!st)
 183                 die("memory allocation failed\n");
 184         st->tid = tid;
 185         INIT_LIST_HEAD(&st->seq_list);
 186
 187         rb_link_node(&st->rb, NULL, &thread_stats.rb_node);
 188         rb_insert_color(&st->rb, &thread_stats);
 189
 190         thread_stat_findnew = thread_stat_findnew_after_first;
 191         return st;
 192 }
 193
 194 /* build simple key function one is bigger than two */
 195 #define SINGLE_KEY(member)                                              \
 196         static int lock_stat_key_ ## member(struct lock_stat *one,      \
 197                                          struct lock_stat *two)         \
 198         {                                                               \
 199                 return one->member > two->member;                       \
 200         }
 201
 202 SINGLE_KEY(nr_acquired)
 203 SINGLE_KEY(nr_contended)
 204 SINGLE_KEY(wait_time_total)
 205 SINGLE_KEY(wait_time_min)
 206 SINGLE_KEY(wait_time_max)
 207
 208 struct lock_key {
 209         /*
 210          * name: the value for specify by user
 211          * this should be simpler than raw name of member
 212          * e.g. nr_acquired -> acquired, wait_time_total -> wait_total
 213          */
 214         const char              *name;
 215         int                     (*key)(struct lock_stat*, struct lock_stat*);
 216 };
 217
 218 static const char               *sort_key = "acquired";
 219
 220 static int                      (*compare)(struct lock_stat *, struct lock_stat *);
 221
 222 static struct rb_root           result; /* place to store sorted data */
 223
 224 #define DEF_KEY_LOCK(name, fn_suffix)   \
 225         { #name, lock_stat_key_ ## fn_suffix }
 226 struct lock_key keys[] = {
 227         DEF_KEY_LOCK(acquired, nr_acquired),
 228         DEF_KEY_LOCK(contended, nr_contended),
 229         DEF_KEY_LOCK(wait_total, wait_time_total),
 230         DEF_KEY_LOCK(wait_min, wait_time_min),
 231         DEF_KEY_LOCK(wait_max, wait_time_max),
 232
 233         /* extra comparisons much complicated should be here */
 234
 235         { NULL, NULL }
 236 };
 237
 238 static void select_key(void)
 239 {
 240         int i;
 241
 242         for (i = 0; keys[i].name; i++) {
 243                 if (!strcmp(keys[i].name, sort_key)) {
 244                         compare = keys[i].key;
 245                         return;
 246                 }
 247         }
 248
 249         die("Unknown compare key:%s\n", sort_key);
 250 }
 251
 252 static void insert_to_result(struct lock_stat *st,
 253                              int (*bigger)(struct lock_stat *, struct lock_stat *))
 254 {
 255         struct rb_node **rb = &result.rb_node;
 256         struct rb_node *parent = NULL;
 257         struct lock_stat *p;
 258
 259         while (*rb) {
 260                 p = container_of(*rb, struct lock_stat, rb);
 261                 parent = *rb;
 262
 263                 if (bigger(st, p))
 264                         rb = &(*rb)->rb_left;
 265                 else
 266                         rb = &(*rb)->rb_right;
 267         }
 268
 269         rb_link_node(&st->rb, parent, rb);
 270         rb_insert_color(&st->rb, &result);
 271 }
 272
 273 /* returns left most element of result, and erase it */
 274 static struct lock_stat *pop_from_result(void)
 275 {
 276         struct rb_node *node = result.rb_node;
 277
 278         if (!node)
 279                 return NULL;
 280
 281         while (node->rb_left)
 282                 node = node->rb_left;
 283
 284         rb_erase(node, &result);
 285         return container_of(node, struct lock_stat, rb);
 286 }
 287
 288 static struct lock_stat *lock_stat_findnew(void *addr, const char *name)
 289 {
 290         struct list_head *entry = lockhashentry(addr);
 291         struct lock_stat *ret, *new;
 292
 293         list_for_each_entry(ret, entry, hash_entry) {
 294                 if (ret->addr == addr)
 295                         return ret;
 296         }
 297
 298         new = zalloc(sizeof(struct lock_stat));
 299         if (!new)
 300                 goto alloc_failed;
 301
 302         new->addr = addr;
 303         new->name = zalloc(sizeof(char) * strlen(name) + 1);
 304         if (!new->name)
 305                 goto alloc_failed;
 306         strcpy(new->name, name);
 307
 308         new->wait_time_min = ULLONG_MAX;
 309
 310         list_add(&new->hash_entry, entry);
 311         return new;
 312
 313 alloc_failed:
 314         die("memory allocation failed\n");
 315 }
 316
 317 static char                     const *input_name = "perf.data";
 318
 319 struct raw_event_sample {
 320         u32                     size;
 321         char                    data[0];
 322 };
 323
 324 struct trace_acquire_event {
 325         void                    *addr;
 326         const char              *name;
 327         int                     flag;
 328 };
 329
 330 struct trace_acquired_event {
 331         void                    *addr;
 332         const char              *name;
 333 };
 334
 335 struct trace_contended_event {
 336         void                    *addr;
 337         const char              *name;
 338 };
 339
 340 struct trace_release_event {
 341         void                    *addr;
 342         const char              *name;
 343 };
 344
 345 struct trace_lock_handler {
 346         void (*acquire_event)(struct trace_acquire_event *,
 347                               struct event *,
 348                               int cpu,
 349                               u64 timestamp,
 350                               struct thread *thread);
 351
 352         void (*acquired_event)(struct trace_acquired_event *,
 353                                struct event *,
 354                                int cpu,
 355                                u64 timestamp,
 356                                struct thread *thread);
 357
 358         void (*contended_event)(struct trace_contended_event *,
 359                                 struct event *,
 360                                 int cpu,
 361                                 u64 timestamp,
 362                                 struct thread *thread);
 363
 364         void (*release_event)(struct trace_release_event *,
 365                               struct event *,
 366                               int cpu,
 367                               u64 timestamp,
 368                               struct thread *thread);
 369 };
 370
 371 static struct lock_seq_stat *get_seq(struct thread_stat *ts, void *addr)
 372 {
 373         struct lock_seq_stat *seq;
 374
 375         list_for_each_entry(seq, &ts->seq_list, list) {
 376                 if (seq->addr == addr)
 377                         return seq;
 378         }
 379
 380         seq = zalloc(sizeof(struct lock_seq_stat));
 381         if (!seq)
 382                 die("Not enough memory\n");
 383         seq->state = SEQ_STATE_UNINITIALIZED;
 384         seq->addr = addr;
 385
 386         list_add(&seq->list, &ts->seq_list);
 387         return seq;
 388 }
 389
 390 enum broken_state {
 391         BROKEN_ACQUIRE,
 392         BROKEN_ACQUIRED,
 393         BROKEN_CONTENDED,
 394         BROKEN_RELEASE,
 395         BROKEN_MAX,
 396 };
 397
 398 static int bad_hist[BROKEN_MAX];
 399
 400 enum acquire_flags {
 401         TRY_LOCK = 1,
 402         READ_LOCK = 2,
 403 };
 404
 405 static void
 406 report_lock_acquire_event(struct trace_acquire_event *acquire_event,
 407                         struct event *__event __used,
 408                         int cpu __used,
 409                         u64 timestamp __used,
 410                         struct thread *thread __used)
 411 {
 412         struct lock_stat *ls;
 413         struct thread_stat *ts;
 414         struct lock_seq_stat *seq;
 415
 416         ls = lock_stat_findnew(acquire_event->addr, acquire_event->name);
 417         if (ls->discard)
 418                 return;
 419
 420         ts = thread_stat_findnew(thread->pid);
 421         seq = get_seq(ts, acquire_event->addr);
 422
 423         switch (seq->state) {
 424         case SEQ_STATE_UNINITIALIZED:
 425         case SEQ_STATE_RELEASED:
 426                 if (!acquire_event->flag) {
 427                         seq->state = SEQ_STATE_ACQUIRING;
 428                 } else {
 429                         if (acquire_event->flag & TRY_LOCK)
 430                                 ls->nr_trylock++;
 431                         if (acquire_event->flag & READ_LOCK)
 432                                 ls->nr_readlock++;
 433                         seq->state = SEQ_STATE_READ_ACQUIRED;
 434                         seq->read_count = 1;
 435                         ls->nr_acquired++;
 436                 }
 437                 break;
 438         case SEQ_STATE_READ_ACQUIRED:
 439                 if (acquire_event->flag & READ_LOCK) {
 440                         seq->read_count++;
 441                         ls->nr_acquired++;
 442                         goto end;
 443                 } else {
 444                         goto broken;
 445                 }
 446                 break;
 447         case SEQ_STATE_ACQUIRED:
 448         case SEQ_STATE_ACQUIRING:
 449         case SEQ_STATE_CONTENDED:
 450 broken:
 451                 /* broken lock sequence, discard it */
 452                 ls->discard = 1;
 453                 bad_hist[BROKEN_ACQUIRE]++;
 454                 list_del(&seq->list);
 455                 free(seq);
 456                 goto end;
 457                 break;
 458         default:
 459                 BUG_ON("Unknown state of lock sequence found!\n");
 460                 break;
 461         }
 462
 463         ls->nr_acquire++;
 464         seq->prev_event_time = timestamp;
 465 end:
 466         return;
 467 }
 468
 469 static void
 470 report_lock_acquired_event(struct trace_acquired_event *acquired_event,
 471                          struct event *__event __used,
 472                          int cpu __used,
 473                          u64 timestamp __used,
 474                          struct thread *thread __used)
 475 {
 476         struct lock_stat *ls;
 477         struct thread_stat *ts;
 478         struct lock_seq_stat *seq;
 479         u64 contended_term;
 480
 481         ls = lock_stat_findnew(acquired_event->addr, acquired_event->name);
 482         if (ls->discard)
 483                 return;
 484
 485         ts = thread_stat_findnew(thread->pid);
 486         seq = get_seq(ts, acquired_event->addr);
 487
 488         switch (seq->state) {
 489         case SEQ_STATE_UNINITIALIZED:
 490                 /* orphan event, do nothing */
 491                 return;
 492         case SEQ_STATE_ACQUIRING:
 493                 break;
 494         case SEQ_STATE_CONTENDED:
 495                 contended_term = timestamp - seq->prev_event_time;
 496                 ls->wait_time_total += contended_term;
 497                 if (contended_term < ls->wait_time_min)
 498                         ls->wait_time_min = contended_term;
 499                 if (ls->wait_time_max < contended_term)
 500                         ls->wait_time_max = contended_term;
 501                 break;
 502         case SEQ_STATE_RELEASED:
 503         case SEQ_STATE_ACQUIRED:
 504         case SEQ_STATE_READ_ACQUIRED:
 505                 /* broken lock sequence, discard it */
 506                 ls->discard = 1;
 507                 bad_hist[BROKEN_ACQUIRED]++;
 508                 list_del(&seq->list);
 509                 free(seq);
 510                 goto end;
 511                 break;
 512
 513         default:
 514                 BUG_ON("Unknown state of lock sequence found!\n");
 515                 break;
 516         }
 517
 518         seq->state = SEQ_STATE_ACQUIRED;
 519         ls->nr_acquired++;
 520         seq->prev_event_time = timestamp;
 521 end:
 522         return;
 523 }
 524
 525 static void
 526 report_lock_contended_event(struct trace_contended_event *contended_event,
 527                           struct event *__event __used,
 528                           int cpu __used,
 529                           u64 timestamp __used,
 530                           struct thread *thread __used)
 531 {
 532         struct lock_stat *ls;
 533         struct thread_stat *ts;
 534         struct lock_seq_stat *seq;
 535
 536         ls = lock_stat_findnew(contended_event->addr, contended_event->name);
 537         if (ls->discard)
 538                 return;
 539
 540         ts = thread_stat_findnew(thread->pid);
 541         seq = get_seq(ts, contended_event->addr);
 542
 543         switch (seq->state) {
 544         case SEQ_STATE_UNINITIALIZED:
 545                 /* orphan event, do nothing */
 546                 return;
 547         case SEQ_STATE_ACQUIRING:
 548                 break;
 549         case SEQ_STATE_RELEASED:
 550         case SEQ_STATE_ACQUIRED:
 551         case SEQ_STATE_READ_ACQUIRED:
 552         case SEQ_STATE_CONTENDED:
 553                 /* broken lock sequence, discard it */
 554                 ls->discard = 1;
 555                 bad_hist[BROKEN_CONTENDED]++;
 556                 list_del(&seq->list);
 557                 free(seq);
 558                 goto end;
 559                 break;
 560         default:
 561                 BUG_ON("Unknown state of lock sequence found!\n");
 562                 break;
 563         }
 564
 565         seq->state = SEQ_STATE_CONTENDED;
 566         ls->nr_contended++;
 567         seq->prev_event_time = timestamp;
 568 end:
 569         return;
 570 }
 571
 572 static void
 573 report_lock_release_event(struct trace_release_event *release_event,
 574                         struct event *__event __used,
 575                         int cpu __used,
 576                         u64 timestamp __used,
 577                         struct thread *thread __used)
 578 {
 579         struct lock_stat *ls;
 580         struct thread_stat *ts;
 581         struct lock_seq_stat *seq;
 582
 583         ls = lock_stat_findnew(release_event->addr, release_event->name);
 584         if (ls->discard)
 585                 return;
 586
 587         ts = thread_stat_findnew(thread->pid);
 588         seq = get_seq(ts, release_event->addr);
 589
 590         switch (seq->state) {
 591         case SEQ_STATE_UNINITIALIZED:
 592                 goto end;
 593                 break;
 594         case SEQ_STATE_ACQUIRED:
 595                 break;
 596         case SEQ_STATE_READ_ACQUIRED:
 597                 seq->read_count--;
 598                 BUG_ON(seq->read_count < 0);
 599                 if (!seq->read_count) {
 600                         ls->nr_release++;
 601                         goto end;
 602                 }
 603                 break;
 604         case SEQ_STATE_ACQUIRING:
 605         case SEQ_STATE_CONTENDED:
 606         case SEQ_STATE_RELEASED:
 607                 /* broken lock sequence, discard it */
 608                 ls->discard = 1;
 609                 bad_hist[BROKEN_RELEASE]++;
 610                 goto free_seq;
 611                 break;
 612         default:
 613                 BUG_ON("Unknown state of lock sequence found!\n");
 614                 break;
 615         }
 616
 617         ls->nr_release++;
 618 free_seq:
 619         list_del(&seq->list);
 620         free(seq);
 621 end:
 622         return;
 623 }
 624
 625 /* lock oriented handlers */
 626 /* TODO: handlers for CPU oriented, thread oriented */
 627 static struct trace_lock_handler report_lock_ops  = {
 628         .acquire_event          = report_lock_acquire_event,
 629         .acquired_event         = report_lock_acquired_event,
 630         .contended_event        = report_lock_contended_event,
 631         .release_event          = report_lock_release_event,
 632 };
 633
 634 static struct trace_lock_handler *trace_handler;
 635
 636 static void
 637 process_lock_acquire_event(void *data,
 638                            struct event *event __used,
 639                            int cpu __used,
 640                            u64 timestamp __used,
 641                            struct thread *thread __used)
 642 {
 643         struct trace_acquire_event acquire_event;
 644         u64 tmp;                /* this is required for casting... */
 645
 646         tmp = raw_field_value(event, "lockdep_addr", data);
 647         memcpy(&acquire_event.addr, &tmp, sizeof(void *));
 648         acquire_event.name = (char *)raw_field_ptr(event, "name", data);
 649         acquire_event.flag = (int)raw_field_value(event, "flag", data);
 650
 651         if (trace_handler->acquire_event)
 652                 trace_handler->acquire_event(&acquire_event, event, cpu, timestamp, thread);
 653 }
 654
 655 static void
 656 process_lock_acquired_event(void *data,
 657                             struct event *event __used,
 658                             int cpu __used,
 659                             u64 timestamp __used,
 660                             struct thread *thread __used)
 661 {
 662         struct trace_acquired_event acquired_event;
 663         u64 tmp;                /* this is required for casting... */
 664
 665         tmp = raw_field_value(event, "lockdep_addr", data);
 666         memcpy(&acquired_event.addr, &tmp, sizeof(void *));
 667         acquired_event.name = (char *)raw_field_ptr(event, "name", data);
 668
 669         if (trace_handler->acquire_event)
 670                 trace_handler->acquired_event(&acquired_event, event, cpu, timestamp, thread);
 671 }
 672
 673 static void
 674 process_lock_contended_event(void *data,
 675                              struct event *event __used,
 676                              int cpu __used,
 677                              u64 timestamp __used,
 678                              struct thread *thread __used)
 679 {
 680         struct trace_contended_event contended_event;
 681         u64 tmp;                /* this is required for casting... */
 682
 683         tmp = raw_field_value(event, "lockdep_addr", data);
 684         memcpy(&contended_event.addr, &tmp, sizeof(void *));
 685         contended_event.name = (char *)raw_field_ptr(event, "name", data);
 686
 687         if (trace_handler->acquire_event)
 688                 trace_handler->contended_event(&contended_event, event, cpu, timestamp, thread);
 689 }
 690
 691 static void
 692 process_lock_release_event(void *data,
 693                            struct event *event __used,
 694                            int cpu __used,
 695                            u64 timestamp __used,
 696                            struct thread *thread __used)
 697 {
 698         struct trace_release_event release_event;
 699         u64 tmp;                /* this is required for casting... */
 700
 701         tmp = raw_field_value(event, "lockdep_addr", data);
 702         memcpy(&release_event.addr, &tmp, sizeof(void *));
 703         release_event.name = (char *)raw_field_ptr(event, "name", data);
 704
 705         if (trace_handler->acquire_event)
 706                 trace_handler->release_event(&release_event, event, cpu, timestamp, thread);
 707 }
 708
 709 static void
 710 process_raw_event(void *data, int cpu, u64 timestamp, struct thread *thread)
 711 {
 712         struct event *event;
 713         int type;
 714
 715         type = trace_parse_common_type(data);
 716         event = trace_find_event(type);
 717
 718         if (!strcmp(event->name, "lock_acquire"))
 719                 process_lock_acquire_event(data, event, cpu, timestamp, thread);
 720         if (!strcmp(event->name, "lock_acquired"))
 721                 process_lock_acquired_event(data, event, cpu, timestamp, thread);
 722         if (!strcmp(event->name, "lock_contended"))
 723                 process_lock_contended_event(data, event, cpu, timestamp, thread);
 724         if (!strcmp(event->name, "lock_release"))
 725                 process_lock_release_event(data, event, cpu, timestamp, thread);
 726 }
 727
 728 static void print_bad_events(int bad, int total)
 729 {
 730         /* Output for debug, this have to be removed */
 731         int i;
 732         const char *name[4] =
 733                 { "acquire", "acquired", "contended", "release" };
 734
 735         pr_info("\n=== output for debug===\n\n");
 736         pr_info("bad: %d, total: %d\n", bad, total);
 737         pr_info("bad rate: %f %%\n", (double)bad / (double)total * 100);
 738         pr_info("histogram of events caused bad sequence\n");
 739         for (i = 0; i < BROKEN_MAX; i++)
 740                 pr_info(" %10s: %d\n", name[i], bad_hist[i]);
 741 }
 742
 743 /* TODO: various way to print, coloring, nano or milli sec */
 744 static void print_result(void)
 745 {
 746         struct lock_stat *st;
 747         char cut_name[20];
 748         int bad, total;
 749
 750         pr_info("%20s ", "Name");
 751         pr_info("%10s ", "acquired");
 752         pr_info("%10s ", "contended");
 753
 754         pr_info("%15s ", "total wait (ns)");
 755         pr_info("%15s ", "max wait (ns)");
 756         pr_info("%15s ", "min wait (ns)");
 757
 758         pr_info("\n\n");
 759
 760         bad = total = 0;
 761         while ((st = pop_from_result())) {
 762                 total++;
 763                 if (st->discard) {
 764                         bad++;
 765                         continue;
 766                 }
 767                 bzero(cut_name, 20);
 768
 769                 if (strlen(st->name) < 16) {
 770                         /* output raw name */
 771                         pr_info("%20s ", st->name);
 772                 } else {
 773                         strncpy(cut_name, st->name, 16);
 774                         cut_name[16] = '.';
 775                         cut_name[17] = '.';
 776                         cut_name[18] = '.';
 777                         cut_name[19] = '\0';
 778                         /* cut off name for saving output style */
 779                         pr_info("%20s ", cut_name);
 780                 }
 781
 782                 pr_info("%10u ", st->nr_acquired);
 783                 pr_info("%10u ", st->nr_contended);
 784
 785                 pr_info("%15" PRIu64 " ", st->wait_time_total);
 786                 pr_info("%15" PRIu64 " ", st->wait_time_max);
 787                 pr_info("%15" PRIu64 " ", st->wait_time_min == ULLONG_MAX ?
 788                        0 : st->wait_time_min);
 789                 pr_info("\n");
 790         }
 791
 792         print_bad_events(bad, total);
 793 }
 794
 795 static bool info_threads, info_map;
 796
 797 static void dump_threads(void)
 798 {
 799         struct thread_stat *st;
 800         struct rb_node *node;
 801         struct thread *t;
 802
 803         pr_info("%10s: comm\n", "Thread ID");
 804
 805         node = rb_first(&thread_stats);
 806         while (node) {
 807                 st = container_of(node, struct thread_stat, rb);
 808                 t = perf_session__findnew(session, st->tid);
 809                 pr_info("%10d: %s\n", st->tid, t->comm);
 810                 node = rb_next(node);
 811         };
 812 }
 813
 814 static void dump_map(void)
 815 {
 816         unsigned int i;
 817         struct lock_stat *st;
 818
 819         pr_info("Address of instance: name of class\n");
 820         for (i = 0; i < LOCKHASH_SIZE; i++) {
 821                 list_for_each_entry(st, &lockhash_table[i], hash_entry) {
 822                         pr_info(" %p: %s\n", st->addr, st->name);
 823                 }
 824         }
 825 }
 826
 827 static void dump_info(void)
 828 {
 829         if (info_threads)
 830                 dump_threads();
 831         else if (info_map)
 832                 dump_map();
 833         else
 834                 die("Unknown type of information\n");
 835 }
 836
 837 static int process_sample_event(event_t *self, struct sample_data *sample,
 838                                 struct perf_session *s)
 839 {
 840         struct thread *thread = perf_session__findnew(s, sample->tid);
 841
 842         if (thread == NULL) {
 843                 pr_debug("problem processing %d event, skipping it.\n",
 844                         self->header.type);
 845                 return -1;
 846         }
 847
 848         process_raw_event(sample->raw_data, sample->cpu, sample->time, thread);
 849
 850         return 0;
 851 }
 852
 853 static struct perf_event_ops eops = {
 854         .sample                 = process_sample_event,
 855         .comm                   = event__process_comm,
 856         .ordered_samples        = true,
 857 };
 858
 859 static int read_events(void)
 860 {
 861         session = perf_session__new(input_name, O_RDONLY, 0, false, &eops);
 862         if (!session)
 863                 die("Initializing perf session failed\n");
 864
 865         return perf_session__process_events(session, &eops);
 866 }
 867
 868 static void sort_result(void)
 869 {
 870         unsigned int i;
 871         struct lock_stat *st;
 872
 873         for (i = 0; i < LOCKHASH_SIZE; i++) {
 874                 list_for_each_entry(st, &lockhash_table[i], hash_entry) {
 875                         insert_to_result(st, compare);
 876                 }
 877         }
 878 }
 879
 880 static void __cmd_report(void)
 881 {
 882         setup_pager();
 883         select_key();
 884         read_events();
 885         sort_result();
 886         print_result();
 887 }
 888
 889 static const char * const report_usage[] = {
 890         "perf lock report [<options>]",
 891         NULL
 892 };
 893
 894 static const struct option report_options[] = {
 895         OPT_STRING('k', "key", &sort_key, "acquired",
 896                     "key for sorting"),
 897         /* TODO: type */
 898         OPT_END()
 899 };
 900
 901 static const char * const info_usage[] = {
 902         "perf lock info [<options>]",
 903         NULL
 904 };
 905
 906 static const struct option info_options[] = {
 907         OPT_BOOLEAN('t', "threads", &info_threads,
 908                     "dump thread list in perf.data"),
 909         OPT_BOOLEAN('m', "map", &info_map,
 910                     "map of lock instances (name:address table)"),
 911         OPT_END()
 912 };
 913
 914 static const char * const lock_usage[] = {
 915         "perf lock [<options>] {record|trace|report}",
 916         NULL
 917 };
 918
 919 static const struct option lock_options[] = {
 920         OPT_STRING('i', "input", &input_name, "file", "input file name"),
 921         OPT_INCR('v', "verbose", &verbose, "be more verbose (show symbol address, etc)"),
 922         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, "dump raw trace in ASCII"),
 923         OPT_END()
 924 };
 925
 926 static const char *record_args[] = {
 927         "record",
 928         "-R",
 929         "-f",
 930         "-m", "1024",
 931         "-c", "1",
 932         "-e", "lock:lock_acquire:r",
 933         "-e", "lock:lock_acquired:r",
 934         "-e", "lock:lock_contended:r",
 935         "-e", "lock:lock_release:r",
 936 };
 937
 938 static int __cmd_record(int argc, const char **argv)
 939 {
 940         unsigned int rec_argc, i, j;
 941         const char **rec_argv;
 942
 943         rec_argc = ARRAY_SIZE(record_args) + argc - 1;
 944         rec_argv = calloc(rec_argc + 1, sizeof(char *));
 945
 946         if (rec_argv == NULL)
 947                 return -ENOMEM;
 948
 949         for (i = 0; i < ARRAY_SIZE(record_args); i++)
 950                 rec_argv[i] = strdup(record_args[i]);
 951
 952         for (j = 1; j < (unsigned int)argc; j++, i++)
 953                 rec_argv[i] = argv[j];
 954
 955         BUG_ON(i != rec_argc);
 956
 957         return cmd_record(i, rec_argv, NULL);
 958 }
 959
 960 int cmd_lock(int argc, const char **argv, const char *prefix __used)
 961 {
 962         unsigned int i;
 963
 964         symbol__init();
 965         for (i = 0; i < LOCKHASH_SIZE; i++)
 966                 INIT_LIST_HEAD(lockhash_table + i);
 967
 968         argc = parse_options(argc, argv, lock_options, lock_usage,
 969                              PARSE_OPT_STOP_AT_NON_OPTION);
 970         if (!argc)
 971                 usage_with_options(lock_usage, lock_options);
 972
 973         if (!strncmp(argv[0], "rec", 3)) {
 974                 return __cmd_record(argc, argv);
 975         } else if (!strncmp(argv[0], "report", 6)) {
 976                 trace_handler = &report_lock_ops;
 977                 if (argc) {
 978                         argc = parse_options(argc, argv,
 979                                              report_options, report_usage, 0);
 980                         if (argc)
 981                                 usage_with_options(report_usage, report_options);
 982                 }
 983                 __cmd_report();
 984         } else if (!strcmp(argv[0], "script")) {
 985                 /* Aliased to 'perf script' */
 986                 return cmd_script(argc, argv, prefix);
 987         } else if (!strcmp(argv[0], "info")) {
 988                 if (argc) {
 989                         argc = parse_options(argc, argv,
 990                                              info_options, info_usage, 0);
 991                         if (argc)
 992                                 usage_with_options(info_usage, info_options);
 993                 }
 994                 /* recycling report_lock_ops */
 995                 trace_handler = &report_lock_ops;
 996                 setup_pager();
 997                 read_events();
 998                 dump_info();
 999         } else {
1000                 usage_with_options(lock_usage, lock_options);
1001         }
1002
1003         return 0;
1004 }