SAFE public projects git trees. - safe/jmp/linux-2.6/blob - tools/perf/builtin-sched.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 #include "util/session.h"
  10
  11 #include "util/parse-options.h"
  12 #include "util/trace-event.h"
  13
  14 #include "util/debug.h"
  15 #include "util/data_map.h"
  16
  17 #include <sys/prctl.h>
  18
  19 #include <semaphore.h>
  20 #include <pthread.h>
  21 #include <math.h>
  22
  23 static char                     const *input_name = "perf.data";
  24
  25 static u64                      sample_type;
  26
  27 static char                     default_sort_order[] = "avg, max, switch, runtime";
  28 static char                     *sort_order = default_sort_order;
  29
  30 static int                      profile_cpu = -1;
  31
  32 #define PR_SET_NAME             15               /* Set process name */
  33 #define MAX_CPUS                4096
  34
  35 static u64                      run_measurement_overhead;
  36 static u64                      sleep_measurement_overhead;
  37
  38 #define COMM_LEN                20
  39 #define SYM_LEN                 129
  40
  41 #define MAX_PID                 65536
  42
  43 static unsigned long            nr_tasks;
  44
  45 struct sched_atom;
  46
  47 struct task_desc {
  48         unsigned long           nr;
  49         unsigned long           pid;
  50         char                    comm[COMM_LEN];
  51
  52         unsigned long           nr_events;
  53         unsigned long           curr_event;
  54         struct sched_atom       **atoms;
  55
  56         pthread_t               thread;
  57         sem_t                   sleep_sem;
  58
  59         sem_t                   ready_for_work;
  60         sem_t                   work_done_sem;
  61
  62         u64                     cpu_usage;
  63 };
  64
  65 enum sched_event_type {
  66         SCHED_EVENT_RUN,
  67         SCHED_EVENT_SLEEP,
  68         SCHED_EVENT_WAKEUP,
  69         SCHED_EVENT_MIGRATION,
  70 };
  71
  72 struct sched_atom {
  73         enum sched_event_type   type;
  74         u64                     timestamp;
  75         u64                     duration;
  76         unsigned long           nr;
  77         int                     specific_wait;
  78         sem_t                   *wait_sem;
  79         struct task_desc        *wakee;
  80 };
  81
  82 static struct task_desc         *pid_to_task[MAX_PID];
  83
  84 static struct task_desc         **tasks;
  85
  86 static pthread_mutex_t          start_work_mutex = PTHREAD_MUTEX_INITIALIZER;
  87 static u64                      start_time;
  88
  89 static pthread_mutex_t          work_done_wait_mutex = PTHREAD_MUTEX_INITIALIZER;
  90
  91 static unsigned long            nr_run_events;
  92 static unsigned long            nr_sleep_events;
  93 static unsigned long            nr_wakeup_events;
  94
  95 static unsigned long            nr_sleep_corrections;
  96 static unsigned long            nr_run_events_optimized;
  97
  98 static unsigned long            targetless_wakeups;
  99 static unsigned long            multitarget_wakeups;
 100
 101 static u64                      cpu_usage;
 102 static u64                      runavg_cpu_usage;
 103 static u64                      parent_cpu_usage;
 104 static u64                      runavg_parent_cpu_usage;
 105
 106 static unsigned long            nr_runs;
 107 static u64                      sum_runtime;
 108 static u64                      sum_fluct;
 109 static u64                      run_avg;
 110
 111 static unsigned long            replay_repeat = 10;
 112 static unsigned long            nr_timestamps;
 113 static unsigned long            nr_unordered_timestamps;
 114 static unsigned long            nr_state_machine_bugs;
 115 static unsigned long            nr_context_switch_bugs;
 116 static unsigned long            nr_events;
 117 static unsigned long            nr_lost_chunks;
 118 static unsigned long            nr_lost_events;
 119
 120 #define TASK_STATE_TO_CHAR_STR "RSDTtZX"
 121
 122 enum thread_state {
 123         THREAD_SLEEPING = 0,
 124         THREAD_WAIT_CPU,
 125         THREAD_SCHED_IN,
 126         THREAD_IGNORE
 127 };
 128
 129 struct work_atom {
 130         struct list_head        list;
 131         enum thread_state       state;
 132         u64                     sched_out_time;
 133         u64                     wake_up_time;
 134         u64                     sched_in_time;
 135         u64                     runtime;
 136 };
 137
 138 struct work_atoms {
 139         struct list_head        work_list;
 140         struct thread           *thread;
 141         struct rb_node          node;
 142         u64                     max_lat;
 143         u64                     max_lat_at;
 144         u64                     total_lat;
 145         u64                     nb_atoms;
 146         u64                     total_runtime;
 147 };
 148
 149 typedef int (*sort_fn_t)(struct work_atoms *, struct work_atoms *);
 150
 151 static struct rb_root           atom_root, sorted_atom_root;
 152
 153 static u64                      all_runtime;
 154 static u64                      all_count;
 155
 156
 157 static u64 get_nsecs(void)
 158 {
 159         struct timespec ts;
 160
 161         clock_gettime(CLOCK_MONOTONIC, &ts);
 162
 163         return ts.tv_sec * 1000000000ULL + ts.tv_nsec;
 164 }
 165
 166 static void burn_nsecs(u64 nsecs)
 167 {
 168         u64 T0 = get_nsecs(), T1;
 169
 170         do {
 171                 T1 = get_nsecs();
 172         } while (T1 + run_measurement_overhead < T0 + nsecs);
 173 }
 174
 175 static void sleep_nsecs(u64 nsecs)
 176 {
 177         struct timespec ts;
 178
 179         ts.tv_nsec = nsecs % 999999999;
 180         ts.tv_sec = nsecs / 999999999;
 181
 182         nanosleep(&ts, NULL);
 183 }
 184
 185 static void calibrate_run_measurement_overhead(void)
 186 {
 187         u64 T0, T1, delta, min_delta = 1000000000ULL;
 188         int i;
 189
 190         for (i = 0; i < 10; i++) {
 191                 T0 = get_nsecs();
 192                 burn_nsecs(0);
 193                 T1 = get_nsecs();
 194                 delta = T1-T0;
 195                 min_delta = min(min_delta, delta);
 196         }
 197         run_measurement_overhead = min_delta;
 198
 199         printf("run measurement overhead: %Ld nsecs\n", min_delta);
 200 }
 201
 202 static void calibrate_sleep_measurement_overhead(void)
 203 {
 204         u64 T0, T1, delta, min_delta = 1000000000ULL;
 205         int i;
 206
 207         for (i = 0; i < 10; i++) {
 208                 T0 = get_nsecs();
 209                 sleep_nsecs(10000);
 210                 T1 = get_nsecs();
 211                 delta = T1-T0;
 212                 min_delta = min(min_delta, delta);
 213         }
 214         min_delta -= 10000;
 215         sleep_measurement_overhead = min_delta;
 216
 217         printf("sleep measurement overhead: %Ld nsecs\n", min_delta);
 218 }
 219
 220 static struct sched_atom *
 221 get_new_event(struct task_desc *task, u64 timestamp)
 222 {
 223         struct sched_atom *event = zalloc(sizeof(*event));
 224         unsigned long idx = task->nr_events;
 225         size_t size;
 226
 227         event->timestamp = timestamp;
 228         event->nr = idx;
 229
 230         task->nr_events++;
 231         size = sizeof(struct sched_atom *) * task->nr_events;
 232         task->atoms = realloc(task->atoms, size);
 233         BUG_ON(!task->atoms);
 234
 235         task->atoms[idx] = event;
 236
 237         return event;
 238 }
 239
 240 static struct sched_atom *last_event(struct task_desc *task)
 241 {
 242         if (!task->nr_events)
 243                 return NULL;
 244
 245         return task->atoms[task->nr_events - 1];
 246 }
 247
 248 static void
 249 add_sched_event_run(struct task_desc *task, u64 timestamp, u64 duration)
 250 {
 251         struct sched_atom *event, *curr_event = last_event(task);
 252
 253         /*
 254          * optimize an existing RUN event by merging this one
 255          * to it:
 256          */
 257         if (curr_event && curr_event->type == SCHED_EVENT_RUN) {
 258                 nr_run_events_optimized++;
 259                 curr_event->duration += duration;
 260                 return;
 261         }
 262
 263         event = get_new_event(task, timestamp);
 264
 265         event->type = SCHED_EVENT_RUN;
 266         event->duration = duration;
 267
 268         nr_run_events++;
 269 }
 270
 271 static void
 272 add_sched_event_wakeup(struct task_desc *task, u64 timestamp,
 273                        struct task_desc *wakee)
 274 {
 275         struct sched_atom *event, *wakee_event;
 276
 277         event = get_new_event(task, timestamp);
 278         event->type = SCHED_EVENT_WAKEUP;
 279         event->wakee = wakee;
 280
 281         wakee_event = last_event(wakee);
 282         if (!wakee_event || wakee_event->type != SCHED_EVENT_SLEEP) {
 283                 targetless_wakeups++;
 284                 return;
 285         }
 286         if (wakee_event->wait_sem) {
 287                 multitarget_wakeups++;
 288                 return;
 289         }
 290
 291         wakee_event->wait_sem = zalloc(sizeof(*wakee_event->wait_sem));
 292         sem_init(wakee_event->wait_sem, 0, 0);
 293         wakee_event->specific_wait = 1;
 294         event->wait_sem = wakee_event->wait_sem;
 295
 296         nr_wakeup_events++;
 297 }
 298
 299 static void
 300 add_sched_event_sleep(struct task_desc *task, u64 timestamp,
 301                       u64 task_state __used)
 302 {
 303         struct sched_atom *event = get_new_event(task, timestamp);
 304
 305         event->type = SCHED_EVENT_SLEEP;
 306
 307         nr_sleep_events++;
 308 }
 309
 310 static struct task_desc *register_pid(unsigned long pid, const char *comm)
 311 {
 312         struct task_desc *task;
 313
 314         BUG_ON(pid >= MAX_PID);
 315
 316         task = pid_to_task[pid];
 317
 318         if (task)
 319                 return task;
 320
 321         task = zalloc(sizeof(*task));
 322         task->pid = pid;
 323         task->nr = nr_tasks;
 324         strcpy(task->comm, comm);
 325         /*
 326          * every task starts in sleeping state - this gets ignored
 327          * if there's no wakeup pointing to this sleep state:
 328          */
 329         add_sched_event_sleep(task, 0, 0);
 330
 331         pid_to_task[pid] = task;
 332         nr_tasks++;
 333         tasks = realloc(tasks, nr_tasks*sizeof(struct task_task *));
 334         BUG_ON(!tasks);
 335         tasks[task->nr] = task;
 336
 337         if (verbose)
 338                 printf("registered task #%ld, PID %ld (%s)\n", nr_tasks, pid, comm);
 339
 340         return task;
 341 }
 342
 343
 344 static void print_task_traces(void)
 345 {
 346         struct task_desc *task;
 347         unsigned long i;
 348
 349         for (i = 0; i < nr_tasks; i++) {
 350                 task = tasks[i];
 351                 printf("task %6ld (%20s:%10ld), nr_events: %ld\n",
 352                         task->nr, task->comm, task->pid, task->nr_events);
 353         }
 354 }
 355
 356 static void add_cross_task_wakeups(void)
 357 {
 358         struct task_desc *task1, *task2;
 359         unsigned long i, j;
 360
 361         for (i = 0; i < nr_tasks; i++) {
 362                 task1 = tasks[i];
 363                 j = i + 1;
 364                 if (j == nr_tasks)
 365                         j = 0;
 366                 task2 = tasks[j];
 367                 add_sched_event_wakeup(task1, 0, task2);
 368         }
 369 }
 370
 371 static void
 372 process_sched_event(struct task_desc *this_task __used, struct sched_atom *atom)
 373 {
 374         int ret = 0;
 375         u64 now;
 376         long long delta;
 377
 378         now = get_nsecs();
 379         delta = start_time + atom->timestamp - now;
 380
 381         switch (atom->type) {
 382                 case SCHED_EVENT_RUN:
 383                         burn_nsecs(atom->duration);
 384                         break;
 385                 case SCHED_EVENT_SLEEP:
 386                         if (atom->wait_sem)
 387                                 ret = sem_wait(atom->wait_sem);
 388                         BUG_ON(ret);
 389                         break;
 390                 case SCHED_EVENT_WAKEUP:
 391                         if (atom->wait_sem)
 392                                 ret = sem_post(atom->wait_sem);
 393                         BUG_ON(ret);
 394                         break;
 395                 case SCHED_EVENT_MIGRATION:
 396                         break;
 397                 default:
 398                         BUG_ON(1);
 399         }
 400 }
 401
 402 static u64 get_cpu_usage_nsec_parent(void)
 403 {
 404         struct rusage ru;
 405         u64 sum;
 406         int err;
 407
 408         err = getrusage(RUSAGE_SELF, &ru);
 409         BUG_ON(err);
 410
 411         sum =  ru.ru_utime.tv_sec*1e9 + ru.ru_utime.tv_usec*1e3;
 412         sum += ru.ru_stime.tv_sec*1e9 + ru.ru_stime.tv_usec*1e3;
 413
 414         return sum;
 415 }
 416
 417 static int self_open_counters(void)
 418 {
 419         struct perf_event_attr attr;
 420         int fd;
 421
 422         memset(&attr, 0, sizeof(attr));
 423
 424         attr.type = PERF_TYPE_SOFTWARE;
 425         attr.config = PERF_COUNT_SW_TASK_CLOCK;
 426
 427         fd = sys_perf_event_open(&attr, 0, -1, -1, 0);
 428
 429         if (fd < 0)
 430                 die("Error: sys_perf_event_open() syscall returned"
 431                     "with %d (%s)\n", fd, strerror(errno));
 432         return fd;
 433 }
 434
 435 static u64 get_cpu_usage_nsec_self(int fd)
 436 {
 437         u64 runtime;
 438         int ret;
 439
 440         ret = read(fd, &runtime, sizeof(runtime));
 441         BUG_ON(ret != sizeof(runtime));
 442
 443         return runtime;
 444 }
 445
 446 static void *thread_func(void *ctx)
 447 {
 448         struct task_desc *this_task = ctx;
 449         u64 cpu_usage_0, cpu_usage_1;
 450         unsigned long i, ret;
 451         char comm2[22];
 452         int fd;
 453
 454         sprintf(comm2, ":%s", this_task->comm);
 455         prctl(PR_SET_NAME, comm2);
 456         fd = self_open_counters();
 457
 458 again:
 459         ret = sem_post(&this_task->ready_for_work);
 460         BUG_ON(ret);
 461         ret = pthread_mutex_lock(&start_work_mutex);
 462         BUG_ON(ret);
 463         ret = pthread_mutex_unlock(&start_work_mutex);
 464         BUG_ON(ret);
 465
 466         cpu_usage_0 = get_cpu_usage_nsec_self(fd);
 467
 468         for (i = 0; i < this_task->nr_events; i++) {
 469                 this_task->curr_event = i;
 470                 process_sched_event(this_task, this_task->atoms[i]);
 471         }
 472
 473         cpu_usage_1 = get_cpu_usage_nsec_self(fd);
 474         this_task->cpu_usage = cpu_usage_1 - cpu_usage_0;
 475         ret = sem_post(&this_task->work_done_sem);
 476         BUG_ON(ret);
 477
 478         ret = pthread_mutex_lock(&work_done_wait_mutex);
 479         BUG_ON(ret);
 480         ret = pthread_mutex_unlock(&work_done_wait_mutex);
 481         BUG_ON(ret);
 482
 483         goto again;
 484 }
 485
 486 static void create_tasks(void)
 487 {
 488         struct task_desc *task;
 489         pthread_attr_t attr;
 490         unsigned long i;
 491         int err;
 492
 493         err = pthread_attr_init(&attr);
 494         BUG_ON(err);
 495         err = pthread_attr_setstacksize(&attr, (size_t)(16*1024));
 496         BUG_ON(err);
 497         err = pthread_mutex_lock(&start_work_mutex);
 498         BUG_ON(err);
 499         err = pthread_mutex_lock(&work_done_wait_mutex);
 500         BUG_ON(err);
 501         for (i = 0; i < nr_tasks; i++) {
 502                 task = tasks[i];
 503                 sem_init(&task->sleep_sem, 0, 0);
 504                 sem_init(&task->ready_for_work, 0, 0);
 505                 sem_init(&task->work_done_sem, 0, 0);
 506                 task->curr_event = 0;
 507                 err = pthread_create(&task->thread, &attr, thread_func, task);
 508                 BUG_ON(err);
 509         }
 510 }
 511
 512 static void wait_for_tasks(void)
 513 {
 514         u64 cpu_usage_0, cpu_usage_1;
 515         struct task_desc *task;
 516         unsigned long i, ret;
 517
 518         start_time = get_nsecs();
 519         cpu_usage = 0;
 520         pthread_mutex_unlock(&work_done_wait_mutex);
 521
 522         for (i = 0; i < nr_tasks; i++) {
 523                 task = tasks[i];
 524                 ret = sem_wait(&task->ready_for_work);
 525                 BUG_ON(ret);
 526                 sem_init(&task->ready_for_work, 0, 0);
 527         }
 528         ret = pthread_mutex_lock(&work_done_wait_mutex);
 529         BUG_ON(ret);
 530
 531         cpu_usage_0 = get_cpu_usage_nsec_parent();
 532
 533         pthread_mutex_unlock(&start_work_mutex);
 534
 535         for (i = 0; i < nr_tasks; i++) {
 536                 task = tasks[i];
 537                 ret = sem_wait(&task->work_done_sem);
 538                 BUG_ON(ret);
 539                 sem_init(&task->work_done_sem, 0, 0);
 540                 cpu_usage += task->cpu_usage;
 541                 task->cpu_usage = 0;
 542         }
 543
 544         cpu_usage_1 = get_cpu_usage_nsec_parent();
 545         if (!runavg_cpu_usage)
 546                 runavg_cpu_usage = cpu_usage;
 547         runavg_cpu_usage = (runavg_cpu_usage*9 + cpu_usage)/10;
 548
 549         parent_cpu_usage = cpu_usage_1 - cpu_usage_0;
 550         if (!runavg_parent_cpu_usage)
 551                 runavg_parent_cpu_usage = parent_cpu_usage;
 552         runavg_parent_cpu_usage = (runavg_parent_cpu_usage*9 +
 553                                    parent_cpu_usage)/10;
 554
 555         ret = pthread_mutex_lock(&start_work_mutex);
 556         BUG_ON(ret);
 557
 558         for (i = 0; i < nr_tasks; i++) {
 559                 task = tasks[i];
 560                 sem_init(&task->sleep_sem, 0, 0);
 561                 task->curr_event = 0;
 562         }
 563 }
 564
 565 static void run_one_test(void)
 566 {
 567         u64 T0, T1, delta, avg_delta, fluct, std_dev;
 568
 569         T0 = get_nsecs();
 570         wait_for_tasks();
 571         T1 = get_nsecs();
 572
 573         delta = T1 - T0;
 574         sum_runtime += delta;
 575         nr_runs++;
 576
 577         avg_delta = sum_runtime / nr_runs;
 578         if (delta < avg_delta)
 579                 fluct = avg_delta - delta;
 580         else
 581                 fluct = delta - avg_delta;
 582         sum_fluct += fluct;
 583         std_dev = sum_fluct / nr_runs / sqrt(nr_runs);
 584         if (!run_avg)
 585                 run_avg = delta;
 586         run_avg = (run_avg*9 + delta)/10;
 587
 588         printf("#%-3ld: %0.3f, ",
 589                 nr_runs, (double)delta/1000000.0);
 590
 591         printf("ravg: %0.2f, ",
 592                 (double)run_avg/1e6);
 593
 594         printf("cpu: %0.2f / %0.2f",
 595                 (double)cpu_usage/1e6, (double)runavg_cpu_usage/1e6);
 596
 597 #if 0
 598         /*
 599          * rusage statistics done by the parent, these are less
 600          * accurate than the sum_exec_runtime based statistics:
 601          */
 602         printf(" [%0.2f / %0.2f]",
 603                 (double)parent_cpu_usage/1e6,
 604                 (double)runavg_parent_cpu_usage/1e6);
 605 #endif
 606
 607         printf("\n");
 608
 609         if (nr_sleep_corrections)
 610                 printf(" (%ld sleep corrections)\n", nr_sleep_corrections);
 611         nr_sleep_corrections = 0;
 612 }
 613
 614 static void test_calibrations(void)
 615 {
 616         u64 T0, T1;
 617
 618         T0 = get_nsecs();
 619         burn_nsecs(1e6);
 620         T1 = get_nsecs();
 621
 622         printf("the run test took %Ld nsecs\n", T1-T0);
 623
 624         T0 = get_nsecs();
 625         sleep_nsecs(1e6);
 626         T1 = get_nsecs();
 627
 628         printf("the sleep test took %Ld nsecs\n", T1-T0);
 629 }
 630
 631 #define FILL_FIELD(ptr, field, event, data)     \
 632         ptr.field = (typeof(ptr.field)) raw_field_value(event, #field, data)
 633
 634 #define FILL_ARRAY(ptr, array, event, data)                     \
 635 do {                                                            \
 636         void *__array = raw_field_ptr(event, #array, data);     \
 637         memcpy(ptr.array, __array, sizeof(ptr.array));  \
 638 } while(0)
 639
 640 #define FILL_COMMON_FIELDS(ptr, event, data)                    \
 641 do {                                                            \
 642         FILL_FIELD(ptr, common_type, event, data);              \
 643         FILL_FIELD(ptr, common_flags, event, data);             \
 644         FILL_FIELD(ptr, common_preempt_count, event, data);     \
 645         FILL_FIELD(ptr, common_pid, event, data);               \
 646         FILL_FIELD(ptr, common_tgid, event, data);              \
 647 } while (0)
 648
 649
 650
 651 struct trace_switch_event {
 652         u32 size;
 653
 654         u16 common_type;
 655         u8 common_flags;
 656         u8 common_preempt_count;
 657         u32 common_pid;
 658         u32 common_tgid;
 659
 660         char prev_comm[16];
 661         u32 prev_pid;
 662         u32 prev_prio;
 663         u64 prev_state;
 664         char next_comm[16];
 665         u32 next_pid;
 666         u32 next_prio;
 667 };
 668
 669 struct trace_runtime_event {
 670         u32 size;
 671
 672         u16 common_type;
 673         u8 common_flags;
 674         u8 common_preempt_count;
 675         u32 common_pid;
 676         u32 common_tgid;
 677
 678         char comm[16];
 679         u32 pid;
 680         u64 runtime;
 681         u64 vruntime;
 682 };
 683
 684 struct trace_wakeup_event {
 685         u32 size;
 686
 687         u16 common_type;
 688         u8 common_flags;
 689         u8 common_preempt_count;
 690         u32 common_pid;
 691         u32 common_tgid;
 692
 693         char comm[16];
 694         u32 pid;
 695
 696         u32 prio;
 697         u32 success;
 698         u32 cpu;
 699 };
 700
 701 struct trace_fork_event {
 702         u32 size;
 703
 704         u16 common_type;
 705         u8 common_flags;
 706         u8 common_preempt_count;
 707         u32 common_pid;
 708         u32 common_tgid;
 709
 710         char parent_comm[16];
 711         u32 parent_pid;
 712         char child_comm[16];
 713         u32 child_pid;
 714 };
 715
 716 struct trace_migrate_task_event {
 717         u32 size;
 718
 719         u16 common_type;
 720         u8 common_flags;
 721         u8 common_preempt_count;
 722         u32 common_pid;
 723         u32 common_tgid;
 724
 725         char comm[16];
 726         u32 pid;
 727
 728         u32 prio;
 729         u32 cpu;
 730 };
 731
 732 struct trace_sched_handler {
 733         void (*switch_event)(struct trace_switch_event *,
 734                              struct event *,
 735                              int cpu,
 736                              u64 timestamp,
 737                              struct thread *thread);
 738
 739         void (*runtime_event)(struct trace_runtime_event *,
 740                               struct event *,
 741                               int cpu,
 742                               u64 timestamp,
 743                               struct thread *thread);
 744
 745         void (*wakeup_event)(struct trace_wakeup_event *,
 746                              struct event *,
 747                              int cpu,
 748                              u64 timestamp,
 749                              struct thread *thread);
 750
 751         void (*fork_event)(struct trace_fork_event *,
 752                            struct event *,
 753                            int cpu,
 754                            u64 timestamp,
 755                            struct thread *thread);
 756
 757         void (*migrate_task_event)(struct trace_migrate_task_event *,
 758                            struct event *,
 759                            int cpu,
 760                            u64 timestamp,
 761                            struct thread *thread);
 762 };
 763
 764
 765 static void
 766 replay_wakeup_event(struct trace_wakeup_event *wakeup_event,
 767                     struct event *event,
 768                     int cpu __used,
 769                     u64 timestamp __used,
 770                     struct thread *thread __used)
 771 {
 772         struct task_desc *waker, *wakee;
 773
 774         if (verbose) {
 775                 printf("sched_wakeup event %p\n", event);
 776
 777                 printf(" ... pid %d woke up %s/%d\n",
 778                         wakeup_event->common_pid,
 779                         wakeup_event->comm,
 780                         wakeup_event->pid);
 781         }
 782
 783         waker = register_pid(wakeup_event->common_pid, "<unknown>");
 784         wakee = register_pid(wakeup_event->pid, wakeup_event->comm);
 785
 786         add_sched_event_wakeup(waker, timestamp, wakee);
 787 }
 788
 789 static u64 cpu_last_switched[MAX_CPUS];
 790
 791 static void
 792 replay_switch_event(struct trace_switch_event *switch_event,
 793                     struct event *event,
 794                     int cpu,
 795                     u64 timestamp,
 796                     struct thread *thread __used)
 797 {
 798         struct task_desc *prev, *next;
 799         u64 timestamp0;
 800         s64 delta;
 801
 802         if (verbose)
 803                 printf("sched_switch event %p\n", event);
 804
 805         if (cpu >= MAX_CPUS || cpu < 0)
 806                 return;
 807
 808         timestamp0 = cpu_last_switched[cpu];
 809         if (timestamp0)
 810                 delta = timestamp - timestamp0;
 811         else
 812                 delta = 0;
 813
 814         if (delta < 0)
 815                 die("hm, delta: %Ld < 0 ?\n", delta);
 816
 817         if (verbose) {
 818                 printf(" ... switch from %s/%d to %s/%d [ran %Ld nsecs]\n",
 819                         switch_event->prev_comm, switch_event->prev_pid,
 820                         switch_event->next_comm, switch_event->next_pid,
 821                         delta);
 822         }
 823
 824         prev = register_pid(switch_event->prev_pid, switch_event->prev_comm);
 825         next = register_pid(switch_event->next_pid, switch_event->next_comm);
 826
 827         cpu_last_switched[cpu] = timestamp;
 828
 829         add_sched_event_run(prev, timestamp, delta);
 830         add_sched_event_sleep(prev, timestamp, switch_event->prev_state);
 831 }
 832
 833
 834 static void
 835 replay_fork_event(struct trace_fork_event *fork_event,
 836                   struct event *event,
 837                   int cpu __used,
 838                   u64 timestamp __used,
 839                   struct thread *thread __used)
 840 {
 841         if (verbose) {
 842                 printf("sched_fork event %p\n", event);
 843                 printf("... parent: %s/%d\n", fork_event->parent_comm, fork_event->parent_pid);
 844                 printf("...  child: %s/%d\n", fork_event->child_comm, fork_event->child_pid);
 845         }
 846         register_pid(fork_event->parent_pid, fork_event->parent_comm);
 847         register_pid(fork_event->child_pid, fork_event->child_comm);
 848 }
 849
 850 static struct trace_sched_handler replay_ops  = {
 851         .wakeup_event           = replay_wakeup_event,
 852         .switch_event           = replay_switch_event,
 853         .fork_event             = replay_fork_event,
 854 };
 855
 856 struct sort_dimension {
 857         const char              *name;
 858         sort_fn_t               cmp;
 859         struct list_head        list;
 860 };
 861
 862 static LIST_HEAD(cmp_pid);
 863
 864 static int
 865 thread_lat_cmp(struct list_head *list, struct work_atoms *l, struct work_atoms *r)
 866 {
 867         struct sort_dimension *sort;
 868         int ret = 0;
 869
 870         BUG_ON(list_empty(list));
 871
 872         list_for_each_entry(sort, list, list) {
 873                 ret = sort->cmp(l, r);
 874                 if (ret)
 875                         return ret;
 876         }
 877
 878         return ret;
 879 }
 880
 881 static struct work_atoms *
 882 thread_atoms_search(struct rb_root *root, struct thread *thread,
 883                          struct list_head *sort_list)
 884 {
 885         struct rb_node *node = root->rb_node;
 886         struct work_atoms key = { .thread = thread };
 887
 888         while (node) {
 889                 struct work_atoms *atoms;
 890                 int cmp;
 891
 892                 atoms = container_of(node, struct work_atoms, node);
 893
 894                 cmp = thread_lat_cmp(sort_list, &key, atoms);
 895                 if (cmp > 0)
 896                         node = node->rb_left;
 897                 else if (cmp < 0)
 898                         node = node->rb_right;
 899                 else {
 900                         BUG_ON(thread != atoms->thread);
 901                         return atoms;
 902                 }
 903         }
 904         return NULL;
 905 }
 906
 907 static void
 908 __thread_latency_insert(struct rb_root *root, struct work_atoms *data,
 909                          struct list_head *sort_list)
 910 {
 911         struct rb_node **new = &(root->rb_node), *parent = NULL;
 912
 913         while (*new) {
 914                 struct work_atoms *this;
 915                 int cmp;
 916
 917                 this = container_of(*new, struct work_atoms, node);
 918                 parent = *new;
 919
 920                 cmp = thread_lat_cmp(sort_list, data, this);
 921
 922                 if (cmp > 0)
 923                         new = &((*new)->rb_left);
 924                 else
 925                         new = &((*new)->rb_right);
 926         }
 927
 928         rb_link_node(&data->node, parent, new);
 929         rb_insert_color(&data->node, root);
 930 }
 931
 932 static void thread_atoms_insert(struct thread *thread)
 933 {
 934         struct work_atoms *atoms = zalloc(sizeof(*atoms));
 935         if (!atoms)
 936                 die("No memory");
 937
 938         atoms->thread = thread;
 939         INIT_LIST_HEAD(&atoms->work_list);
 940         __thread_latency_insert(&atom_root, atoms, &cmp_pid);
 941 }
 942
 943 static void
 944 latency_fork_event(struct trace_fork_event *fork_event __used,
 945                    struct event *event __used,
 946                    int cpu __used,
 947                    u64 timestamp __used,
 948                    struct thread *thread __used)
 949 {
 950         /* should insert the newcomer */
 951 }
 952
 953 __used
 954 static char sched_out_state(struct trace_switch_event *switch_event)
 955 {
 956         const char *str = TASK_STATE_TO_CHAR_STR;
 957
 958         return str[switch_event->prev_state];
 959 }
 960
 961 static void
 962 add_sched_out_event(struct work_atoms *atoms,
 963                     char run_state,
 964                     u64 timestamp)
 965 {
 966         struct work_atom *atom = zalloc(sizeof(*atom));
 967         if (!atom)
 968                 die("Non memory");
 969
 970         atom->sched_out_time = timestamp;
 971
 972         if (run_state == 'R') {
 973                 atom->state = THREAD_WAIT_CPU;
 974                 atom->wake_up_time = atom->sched_out_time;
 975         }
 976
 977         list_add_tail(&atom->list, &atoms->work_list);
 978 }
 979
 980 static void
 981 add_runtime_event(struct work_atoms *atoms, u64 delta, u64 timestamp __used)
 982 {
 983         struct work_atom *atom;
 984
 985         BUG_ON(list_empty(&atoms->work_list));
 986
 987         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
 988
 989         atom->runtime += delta;
 990         atoms->total_runtime += delta;
 991 }
 992
 993 static void
 994 add_sched_in_event(struct work_atoms *atoms, u64 timestamp)
 995 {
 996         struct work_atom *atom;
 997         u64 delta;
 998
 999         if (list_empty(&atoms->work_list))
1000                 return;
1001
1002         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1003
1004         if (atom->state != THREAD_WAIT_CPU)
1005                 return;
1006
1007         if (timestamp < atom->wake_up_time) {
1008                 atom->state = THREAD_IGNORE;
1009                 return;
1010         }
1011
1012         atom->state = THREAD_SCHED_IN;
1013         atom->sched_in_time = timestamp;
1014
1015         delta = atom->sched_in_time - atom->wake_up_time;
1016         atoms->total_lat += delta;
1017         if (delta > atoms->max_lat) {
1018                 atoms->max_lat = delta;
1019                 atoms->max_lat_at = timestamp;
1020         }
1021         atoms->nb_atoms++;
1022 }
1023
1024 static void
1025 latency_switch_event(struct trace_switch_event *switch_event,
1026                      struct event *event __used,
1027                      int cpu,
1028                      u64 timestamp,
1029                      struct thread *thread __used)
1030 {
1031         struct work_atoms *out_events, *in_events;
1032         struct thread *sched_out, *sched_in;
1033         u64 timestamp0;
1034         s64 delta;
1035
1036         BUG_ON(cpu >= MAX_CPUS || cpu < 0);
1037
1038         timestamp0 = cpu_last_switched[cpu];
1039         cpu_last_switched[cpu] = timestamp;
1040         if (timestamp0)
1041                 delta = timestamp - timestamp0;
1042         else
1043                 delta = 0;
1044
1045         if (delta < 0)
1046                 die("hm, delta: %Ld < 0 ?\n", delta);
1047
1048
1049         sched_out = threads__findnew(switch_event->prev_pid);
1050         sched_in = threads__findnew(switch_event->next_pid);
1051
1052         out_events = thread_atoms_search(&atom_root, sched_out, &cmp_pid);
1053         if (!out_events) {
1054                 thread_atoms_insert(sched_out);
1055                 out_events = thread_atoms_search(&atom_root, sched_out, &cmp_pid);
1056                 if (!out_events)
1057                         die("out-event: Internal tree error");
1058         }
1059         add_sched_out_event(out_events, sched_out_state(switch_event), timestamp);
1060
1061         in_events = thread_atoms_search(&atom_root, sched_in, &cmp_pid);
1062         if (!in_events) {
1063                 thread_atoms_insert(sched_in);
1064                 in_events = thread_atoms_search(&atom_root, sched_in, &cmp_pid);
1065                 if (!in_events)
1066                         die("in-event: Internal tree error");
1067                 /*
1068                  * Take came in we have not heard about yet,
1069                  * add in an initial atom in runnable state:
1070                  */
1071                 add_sched_out_event(in_events, 'R', timestamp);
1072         }
1073         add_sched_in_event(in_events, timestamp);
1074 }
1075
1076 static void
1077 latency_runtime_event(struct trace_runtime_event *runtime_event,
1078                      struct event *event __used,
1079                      int cpu,
1080                      u64 timestamp,
1081                      struct thread *this_thread __used)
1082 {
1083         struct thread *thread = threads__findnew(runtime_event->pid);
1084         struct work_atoms *atoms = thread_atoms_search(&atom_root, thread, &cmp_pid);
1085
1086         BUG_ON(cpu >= MAX_CPUS || cpu < 0);
1087         if (!atoms) {
1088                 thread_atoms_insert(thread);
1089                 atoms = thread_atoms_search(&atom_root, thread, &cmp_pid);
1090                 if (!atoms)
1091                         die("in-event: Internal tree error");
1092                 add_sched_out_event(atoms, 'R', timestamp);
1093         }
1094
1095         add_runtime_event(atoms, runtime_event->runtime, timestamp);
1096 }
1097
1098 static void
1099 latency_wakeup_event(struct trace_wakeup_event *wakeup_event,
1100                      struct event *__event __used,
1101                      int cpu __used,
1102                      u64 timestamp,
1103                      struct thread *thread __used)
1104 {
1105         struct work_atoms *atoms;
1106         struct work_atom *atom;
1107         struct thread *wakee;
1108
1109         /* Note for later, it may be interesting to observe the failing cases */
1110         if (!wakeup_event->success)
1111                 return;
1112
1113         wakee = threads__findnew(wakeup_event->pid);
1114         atoms = thread_atoms_search(&atom_root, wakee, &cmp_pid);
1115         if (!atoms) {
1116                 thread_atoms_insert(wakee);
1117                 atoms = thread_atoms_search(&atom_root, wakee, &cmp_pid);
1118                 if (!atoms)
1119                         die("wakeup-event: Internal tree error");
1120                 add_sched_out_event(atoms, 'S', timestamp);
1121         }
1122
1123         BUG_ON(list_empty(&atoms->work_list));
1124
1125         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1126
1127         /*
1128          * You WILL be missing events if you've recorded only
1129          * one CPU, or are only looking at only one, so don't
1130          * make useless noise.
1131          */
1132         if (profile_cpu == -1 && atom->state != THREAD_SLEEPING)
1133                 nr_state_machine_bugs++;
1134
1135         nr_timestamps++;
1136         if (atom->sched_out_time > timestamp) {
1137                 nr_unordered_timestamps++;
1138                 return;
1139         }
1140
1141         atom->state = THREAD_WAIT_CPU;
1142         atom->wake_up_time = timestamp;
1143 }
1144
1145 static void
1146 latency_migrate_task_event(struct trace_migrate_task_event *migrate_task_event,
1147                      struct event *__event __used,
1148                      int cpu __used,
1149                      u64 timestamp,
1150                      struct thread *thread __used)
1151 {
1152         struct work_atoms *atoms;
1153         struct work_atom *atom;
1154         struct thread *migrant;
1155
1156         /*
1157          * Only need to worry about migration when profiling one CPU.
1158          */
1159         if (profile_cpu == -1)
1160                 return;
1161
1162         migrant = threads__findnew(migrate_task_event->pid);
1163         atoms = thread_atoms_search(&atom_root, migrant, &cmp_pid);
1164         if (!atoms) {
1165                 thread_atoms_insert(migrant);
1166                 register_pid(migrant->pid, migrant->comm);
1167                 atoms = thread_atoms_search(&atom_root, migrant, &cmp_pid);
1168                 if (!atoms)
1169                         die("migration-event: Internal tree error");
1170                 add_sched_out_event(atoms, 'R', timestamp);
1171         }
1172
1173         BUG_ON(list_empty(&atoms->work_list));
1174
1175         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1176         atom->sched_in_time = atom->sched_out_time = atom->wake_up_time = timestamp;
1177
1178         nr_timestamps++;
1179
1180         if (atom->sched_out_time > timestamp)
1181                 nr_unordered_timestamps++;
1182 }
1183
1184 static struct trace_sched_handler lat_ops  = {
1185         .wakeup_event           = latency_wakeup_event,
1186         .switch_event           = latency_switch_event,
1187         .runtime_event          = latency_runtime_event,
1188         .fork_event             = latency_fork_event,
1189         .migrate_task_event     = latency_migrate_task_event,
1190 };
1191
1192 static void output_lat_thread(struct work_atoms *work_list)
1193 {
1194         int i;
1195         int ret;
1196         u64 avg;
1197
1198         if (!work_list->nb_atoms)
1199                 return;
1200         /*
1201          * Ignore idle threads:
1202          */
1203         if (!strcmp(work_list->thread->comm, "swapper"))
1204                 return;
1205
1206         all_runtime += work_list->total_runtime;
1207         all_count += work_list->nb_atoms;
1208
1209         ret = printf("  %s:%d ", work_list->thread->comm, work_list->thread->pid);
1210
1211         for (i = 0; i < 24 - ret; i++)
1212                 printf(" ");
1213
1214         avg = work_list->total_lat / work_list->nb_atoms;
1215
1216         printf("|%11.3f ms |%9llu | avg:%9.3f ms | max:%9.3f ms | max at: %9.6f s\n",
1217               (double)work_list->total_runtime / 1e6,
1218                  work_list->nb_atoms, (double)avg / 1e6,
1219                  (double)work_list->max_lat / 1e6,
1220                  (double)work_list->max_lat_at / 1e9);
1221 }
1222
1223 static int pid_cmp(struct work_atoms *l, struct work_atoms *r)
1224 {
1225         if (l->thread->pid < r->thread->pid)
1226                 return -1;
1227         if (l->thread->pid > r->thread->pid)
1228                 return 1;
1229
1230         return 0;
1231 }
1232
1233 static struct sort_dimension pid_sort_dimension = {
1234         .name                   = "pid",
1235         .cmp                    = pid_cmp,
1236 };
1237
1238 static int avg_cmp(struct work_atoms *l, struct work_atoms *r)
1239 {
1240         u64 avgl, avgr;
1241
1242         if (!l->nb_atoms)
1243                 return -1;
1244
1245         if (!r->nb_atoms)
1246                 return 1;
1247
1248         avgl = l->total_lat / l->nb_atoms;
1249         avgr = r->total_lat / r->nb_atoms;
1250
1251         if (avgl < avgr)
1252                 return -1;
1253         if (avgl > avgr)
1254                 return 1;
1255
1256         return 0;
1257 }
1258
1259 static struct sort_dimension avg_sort_dimension = {
1260         .name                   = "avg",
1261         .cmp                    = avg_cmp,
1262 };
1263
1264 static int max_cmp(struct work_atoms *l, struct work_atoms *r)
1265 {
1266         if (l->max_lat < r->max_lat)
1267                 return -1;
1268         if (l->max_lat > r->max_lat)
1269                 return 1;
1270
1271         return 0;
1272 }
1273
1274 static struct sort_dimension max_sort_dimension = {
1275         .name                   = "max",
1276         .cmp                    = max_cmp,
1277 };
1278
1279 static int switch_cmp(struct work_atoms *l, struct work_atoms *r)
1280 {
1281         if (l->nb_atoms < r->nb_atoms)
1282                 return -1;
1283         if (l->nb_atoms > r->nb_atoms)
1284                 return 1;
1285
1286         return 0;
1287 }
1288
1289 static struct sort_dimension switch_sort_dimension = {
1290         .name                   = "switch",
1291         .cmp                    = switch_cmp,
1292 };
1293
1294 static int runtime_cmp(struct work_atoms *l, struct work_atoms *r)
1295 {
1296         if (l->total_runtime < r->total_runtime)
1297                 return -1;
1298         if (l->total_runtime > r->total_runtime)
1299                 return 1;
1300
1301         return 0;
1302 }
1303
1304 static struct sort_dimension runtime_sort_dimension = {
1305         .name                   = "runtime",
1306         .cmp                    = runtime_cmp,
1307 };
1308
1309 static struct sort_dimension *available_sorts[] = {
1310         &pid_sort_dimension,
1311         &avg_sort_dimension,
1312         &max_sort_dimension,
1313         &switch_sort_dimension,
1314         &runtime_sort_dimension,
1315 };
1316
1317 #define NB_AVAILABLE_SORTS      (int)(sizeof(available_sorts) / sizeof(struct sort_dimension *))
1318
1319 static LIST_HEAD(sort_list);
1320
1321 static int sort_dimension__add(const char *tok, struct list_head *list)
1322 {
1323         int i;
1324
1325         for (i = 0; i < NB_AVAILABLE_SORTS; i++) {
1326                 if (!strcmp(available_sorts[i]->name, tok)) {
1327                         list_add_tail(&available_sorts[i]->list, list);
1328
1329                         return 0;
1330                 }
1331         }
1332
1333         return -1;
1334 }
1335
1336 static void setup_sorting(void);
1337
1338 static void sort_lat(void)
1339 {
1340         struct rb_node *node;
1341
1342         for (;;) {
1343                 struct work_atoms *data;
1344                 node = rb_first(&atom_root);
1345                 if (!node)
1346                         break;
1347
1348                 rb_erase(node, &atom_root);
1349                 data = rb_entry(node, struct work_atoms, node);
1350                 __thread_latency_insert(&sorted_atom_root, data, &sort_list);
1351         }
1352 }
1353
1354 static struct trace_sched_handler *trace_handler;
1355
1356 static void
1357 process_sched_wakeup_event(void *data,
1358                            struct event *event,
1359                            int cpu __used,
1360                            u64 timestamp __used,
1361                            struct thread *thread __used)
1362 {
1363         struct trace_wakeup_event wakeup_event;
1364
1365         FILL_COMMON_FIELDS(wakeup_event, event, data);
1366
1367         FILL_ARRAY(wakeup_event, comm, event, data);
1368         FILL_FIELD(wakeup_event, pid, event, data);
1369         FILL_FIELD(wakeup_event, prio, event, data);
1370         FILL_FIELD(wakeup_event, success, event, data);
1371         FILL_FIELD(wakeup_event, cpu, event, data);
1372
1373         if (trace_handler->wakeup_event)
1374                 trace_handler->wakeup_event(&wakeup_event, event, cpu, timestamp, thread);
1375 }
1376
1377 /*
1378  * Track the current task - that way we can know whether there's any
1379  * weird events, such as a task being switched away that is not current.
1380  */
1381 static int max_cpu;
1382
1383 static u32 curr_pid[MAX_CPUS] = { [0 ... MAX_CPUS-1] = -1 };
1384
1385 static struct thread *curr_thread[MAX_CPUS];
1386
1387 static char next_shortname1 = 'A';
1388 static char next_shortname2 = '0';
1389
1390 static void
1391 map_switch_event(struct trace_switch_event *switch_event,
1392                  struct event *event __used,
1393                  int this_cpu,
1394                  u64 timestamp,
1395                  struct thread *thread __used)
1396 {
1397         struct thread *sched_out, *sched_in;
1398         int new_shortname;
1399         u64 timestamp0;
1400         s64 delta;
1401         int cpu;
1402
1403         BUG_ON(this_cpu >= MAX_CPUS || this_cpu < 0);
1404
1405         if (this_cpu > max_cpu)
1406                 max_cpu = this_cpu;
1407
1408         timestamp0 = cpu_last_switched[this_cpu];
1409         cpu_last_switched[this_cpu] = timestamp;
1410         if (timestamp0)
1411                 delta = timestamp - timestamp0;
1412         else
1413                 delta = 0;
1414
1415         if (delta < 0)
1416                 die("hm, delta: %Ld < 0 ?\n", delta);
1417
1418
1419         sched_out = threads__findnew(switch_event->prev_pid);
1420         sched_in = threads__findnew(switch_event->next_pid);
1421
1422         curr_thread[this_cpu] = sched_in;
1423
1424         printf("  ");
1425
1426         new_shortname = 0;
1427         if (!sched_in->shortname[0]) {
1428                 sched_in->shortname[0] = next_shortname1;
1429                 sched_in->shortname[1] = next_shortname2;
1430
1431                 if (next_shortname1 < 'Z') {
1432                         next_shortname1++;
1433                 } else {
1434                         next_shortname1='A';
1435                         if (next_shortname2 < '9') {
1436                                 next_shortname2++;
1437                         } else {
1438                                 next_shortname2='0';
1439                         }
1440                 }
1441                 new_shortname = 1;
1442         }
1443
1444         for (cpu = 0; cpu <= max_cpu; cpu++) {
1445                 if (cpu != this_cpu)
1446                         printf(" ");
1447                 else
1448                         printf("*");
1449
1450                 if (curr_thread[cpu]) {
1451                         if (curr_thread[cpu]->pid)
1452                                 printf("%2s ", curr_thread[cpu]->shortname);
1453                         else
1454                                 printf(".  ");
1455                 } else
1456                         printf("   ");
1457         }
1458
1459         printf("  %12.6f secs ", (double)timestamp/1e9);
1460         if (new_shortname) {
1461                 printf("%s => %s:%d\n",
1462                         sched_in->shortname, sched_in->comm, sched_in->pid);
1463         } else {
1464                 printf("\n");
1465         }
1466 }
1467
1468
1469 static void
1470 process_sched_switch_event(void *data,
1471                            struct event *event,
1472                            int this_cpu,
1473                            u64 timestamp __used,
1474                            struct thread *thread __used)
1475 {
1476         struct trace_switch_event switch_event;
1477
1478         FILL_COMMON_FIELDS(switch_event, event, data);
1479
1480         FILL_ARRAY(switch_event, prev_comm, event, data);
1481         FILL_FIELD(switch_event, prev_pid, event, data);
1482         FILL_FIELD(switch_event, prev_prio, event, data);
1483         FILL_FIELD(switch_event, prev_state, event, data);
1484         FILL_ARRAY(switch_event, next_comm, event, data);
1485         FILL_FIELD(switch_event, next_pid, event, data);
1486         FILL_FIELD(switch_event, next_prio, event, data);
1487
1488         if (curr_pid[this_cpu] != (u32)-1) {
1489                 /*
1490                  * Are we trying to switch away a PID that is
1491                  * not current?
1492                  */
1493                 if (curr_pid[this_cpu] != switch_event.prev_pid)
1494                         nr_context_switch_bugs++;
1495         }
1496         if (trace_handler->switch_event)
1497                 trace_handler->switch_event(&switch_event, event, this_cpu, timestamp, thread);
1498
1499         curr_pid[this_cpu] = switch_event.next_pid;
1500 }
1501
1502 static void
1503 process_sched_runtime_event(void *data,
1504                            struct event *event,
1505                            int cpu __used,
1506                            u64 timestamp __used,
1507                            struct thread *thread __used)
1508 {
1509         struct trace_runtime_event runtime_event;
1510
1511         FILL_ARRAY(runtime_event, comm, event, data);
1512         FILL_FIELD(runtime_event, pid, event, data);
1513         FILL_FIELD(runtime_event, runtime, event, data);
1514         FILL_FIELD(runtime_event, vruntime, event, data);
1515
1516         if (trace_handler->runtime_event)
1517                 trace_handler->runtime_event(&runtime_event, event, cpu, timestamp, thread);
1518 }
1519
1520 static void
1521 process_sched_fork_event(void *data,
1522                          struct event *event,
1523                          int cpu __used,
1524                          u64 timestamp __used,
1525                          struct thread *thread __used)
1526 {
1527         struct trace_fork_event fork_event;
1528
1529         FILL_COMMON_FIELDS(fork_event, event, data);
1530
1531         FILL_ARRAY(fork_event, parent_comm, event, data);
1532         FILL_FIELD(fork_event, parent_pid, event, data);
1533         FILL_ARRAY(fork_event, child_comm, event, data);
1534         FILL_FIELD(fork_event, child_pid, event, data);
1535
1536         if (trace_handler->fork_event)
1537                 trace_handler->fork_event(&fork_event, event, cpu, timestamp, thread);
1538 }
1539
1540 static void
1541 process_sched_exit_event(struct event *event,
1542                          int cpu __used,
1543                          u64 timestamp __used,
1544                          struct thread *thread __used)
1545 {
1546         if (verbose)
1547                 printf("sched_exit event %p\n", event);
1548 }
1549
1550 static void
1551 process_sched_migrate_task_event(void *data,
1552                            struct event *event,
1553                            int cpu __used,
1554                            u64 timestamp __used,
1555                            struct thread *thread __used)
1556 {
1557         struct trace_migrate_task_event migrate_task_event;
1558
1559         FILL_COMMON_FIELDS(migrate_task_event, event, data);
1560
1561         FILL_ARRAY(migrate_task_event, comm, event, data);
1562         FILL_FIELD(migrate_task_event, pid, event, data);
1563         FILL_FIELD(migrate_task_event, prio, event, data);
1564         FILL_FIELD(migrate_task_event, cpu, event, data);
1565
1566         if (trace_handler->migrate_task_event)
1567                 trace_handler->migrate_task_event(&migrate_task_event, event, cpu, timestamp, thread);
1568 }
1569
1570 static void
1571 process_raw_event(event_t *raw_event __used, void *data,
1572                   int cpu, u64 timestamp, struct thread *thread)
1573 {
1574         struct event *event;
1575         int type;
1576
1577
1578         type = trace_parse_common_type(data);
1579         event = trace_find_event(type);
1580
1581         if (!strcmp(event->name, "sched_switch"))
1582                 process_sched_switch_event(data, event, cpu, timestamp, thread);
1583         if (!strcmp(event->name, "sched_stat_runtime"))
1584                 process_sched_runtime_event(data, event, cpu, timestamp, thread);
1585         if (!strcmp(event->name, "sched_wakeup"))
1586                 process_sched_wakeup_event(data, event, cpu, timestamp, thread);
1587         if (!strcmp(event->name, "sched_wakeup_new"))
1588                 process_sched_wakeup_event(data, event, cpu, timestamp, thread);
1589         if (!strcmp(event->name, "sched_process_fork"))
1590                 process_sched_fork_event(data, event, cpu, timestamp, thread);
1591         if (!strcmp(event->name, "sched_process_exit"))
1592                 process_sched_exit_event(event, cpu, timestamp, thread);
1593         if (!strcmp(event->name, "sched_migrate_task"))
1594                 process_sched_migrate_task_event(data, event, cpu, timestamp, thread);
1595 }
1596
1597 static int process_sample_event(event_t *event,
1598                                 struct perf_session *session __used)
1599 {
1600         struct sample_data data;
1601         struct thread *thread;
1602
1603         if (!(sample_type & PERF_SAMPLE_RAW))
1604                 return 0;
1605
1606         memset(&data, 0, sizeof(data));
1607         data.time = -1;
1608         data.cpu = -1;
1609         data.period = -1;
1610
1611         event__parse_sample(event, sample_type, &data);
1612
1613         dump_printf("(IP, %d): %d/%d: %p period: %Ld\n",
1614                 event->header.misc,
1615                 data.pid, data.tid,
1616                 (void *)(long)data.ip,
1617                 (long long)data.period);
1618
1619         thread = threads__findnew(data.pid);
1620         if (thread == NULL) {
1621                 pr_debug("problem processing %d event, skipping it.\n",
1622                          event->header.type);
1623                 return -1;
1624         }
1625
1626         dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid);
1627
1628         if (profile_cpu != -1 && profile_cpu != (int)data.cpu)
1629                 return 0;
1630
1631         process_raw_event(event, data.raw_data, data.cpu, data.time, thread);
1632
1633         return 0;
1634 }
1635
1636 static int process_lost_event(event_t *event __used,
1637                               struct perf_session *session __used)
1638 {
1639         nr_lost_chunks++;
1640         nr_lost_events += event->lost.lost;
1641
1642         return 0;
1643 }
1644
1645 static int sample_type_check(u64 type)
1646 {
1647         sample_type = type;
1648
1649         if (!(sample_type & PERF_SAMPLE_RAW)) {
1650                 fprintf(stderr,
1651                         "No trace sample to read. Did you call perf record "
1652                         "without -R?");
1653                 return -1;
1654         }
1655
1656         return 0;
1657 }
1658
1659 static struct perf_file_handler file_handler = {
1660         .process_sample_event   = process_sample_event,
1661         .process_comm_event     = event__process_comm,
1662         .process_lost_event     = process_lost_event,
1663         .sample_type_check      = sample_type_check,
1664 };
1665
1666 static int read_events(void)
1667 {
1668         int err;
1669         struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0);
1670
1671         if (session == NULL)
1672                 return -ENOMEM;
1673
1674         register_idle_thread();
1675         register_perf_file_handler(&file_handler);
1676
1677         err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd);
1678         perf_session__delete(session);
1679         return err;
1680 }
1681
1682 static void print_bad_events(void)
1683 {
1684         if (nr_unordered_timestamps && nr_timestamps) {
1685                 printf("  INFO: %.3f%% unordered timestamps (%ld out of %ld)\n",
1686                         (double)nr_unordered_timestamps/(double)nr_timestamps*100.0,
1687                         nr_unordered_timestamps, nr_timestamps);
1688         }
1689         if (nr_lost_events && nr_events) {
1690                 printf("  INFO: %.3f%% lost events (%ld out of %ld, in %ld chunks)\n",
1691                         (double)nr_lost_events/(double)nr_events*100.0,
1692                         nr_lost_events, nr_events, nr_lost_chunks);
1693         }
1694         if (nr_state_machine_bugs && nr_timestamps) {
1695                 printf("  INFO: %.3f%% state machine bugs (%ld out of %ld)",
1696                         (double)nr_state_machine_bugs/(double)nr_timestamps*100.0,
1697                         nr_state_machine_bugs, nr_timestamps);
1698                 if (nr_lost_events)
1699                         printf(" (due to lost events?)");
1700                 printf("\n");
1701         }
1702         if (nr_context_switch_bugs && nr_timestamps) {
1703                 printf("  INFO: %.3f%% context switch bugs (%ld out of %ld)",
1704                         (double)nr_context_switch_bugs/(double)nr_timestamps*100.0,
1705                         nr_context_switch_bugs, nr_timestamps);
1706                 if (nr_lost_events)
1707                         printf(" (due to lost events?)");
1708                 printf("\n");
1709         }
1710 }
1711
1712 static void __cmd_lat(void)
1713 {
1714         struct rb_node *next;
1715
1716         setup_pager();
1717         read_events();
1718         sort_lat();
1719
1720         printf("\n ---------------------------------------------------------------------------------------------------------------\n");
1721         printf("  Task                  |   Runtime ms  | Switches | Average delay ms | Maximum delay ms | Maximum delay at     |\n");
1722         printf(" ---------------------------------------------------------------------------------------------------------------\n");
1723
1724         next = rb_first(&sorted_atom_root);
1725
1726         while (next) {
1727                 struct work_atoms *work_list;
1728
1729                 work_list = rb_entry(next, struct work_atoms, node);
1730                 output_lat_thread(work_list);
1731                 next = rb_next(next);
1732         }
1733
1734         printf(" -----------------------------------------------------------------------------------------\n");
1735         printf("  TOTAL:                |%11.3f ms |%9Ld |\n",
1736                 (double)all_runtime/1e6, all_count);
1737
1738         printf(" ---------------------------------------------------\n");
1739
1740         print_bad_events();
1741         printf("\n");
1742
1743 }
1744
1745 static struct trace_sched_handler map_ops  = {
1746         .wakeup_event           = NULL,
1747         .switch_event           = map_switch_event,
1748         .runtime_event          = NULL,
1749         .fork_event             = NULL,
1750 };
1751
1752 static void __cmd_map(void)
1753 {
1754         max_cpu = sysconf(_SC_NPROCESSORS_CONF);
1755
1756         setup_pager();
1757         read_events();
1758         print_bad_events();
1759 }
1760
1761 static void __cmd_replay(void)
1762 {
1763         unsigned long i;
1764
1765         calibrate_run_measurement_overhead();
1766         calibrate_sleep_measurement_overhead();
1767
1768         test_calibrations();
1769
1770         read_events();
1771
1772         printf("nr_run_events:        %ld\n", nr_run_events);
1773         printf("nr_sleep_events:      %ld\n", nr_sleep_events);
1774         printf("nr_wakeup_events:     %ld\n", nr_wakeup_events);
1775
1776         if (targetless_wakeups)
1777                 printf("target-less wakeups:  %ld\n", targetless_wakeups);
1778         if (multitarget_wakeups)
1779                 printf("multi-target wakeups: %ld\n", multitarget_wakeups);
1780         if (nr_run_events_optimized)
1781                 printf("run atoms optimized: %ld\n",
1782                         nr_run_events_optimized);
1783
1784         print_task_traces();
1785         add_cross_task_wakeups();
1786
1787         create_tasks();
1788         printf("------------------------------------------------------------\n");
1789         for (i = 0; i < replay_repeat; i++)
1790                 run_one_test();
1791 }
1792
1793
1794 static const char * const sched_usage[] = {
1795         "perf sched [<options>] {record|latency|map|replay|trace}",
1796         NULL
1797 };
1798
1799 static const struct option sched_options[] = {
1800         OPT_STRING('i', "input", &input_name, "file",
1801                     "input file name"),
1802         OPT_BOOLEAN('v', "verbose", &verbose,
1803                     "be more verbose (show symbol address, etc)"),
1804         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
1805                     "dump raw trace in ASCII"),
1806         OPT_END()
1807 };
1808
1809 static const char * const latency_usage[] = {
1810         "perf sched latency [<options>]",
1811         NULL
1812 };
1813
1814 static const struct option latency_options[] = {
1815         OPT_STRING('s', "sort", &sort_order, "key[,key2...]",
1816                    "sort by key(s): runtime, switch, avg, max"),
1817         OPT_BOOLEAN('v', "verbose", &verbose,
1818                     "be more verbose (show symbol address, etc)"),
1819         OPT_INTEGER('C', "CPU", &profile_cpu,
1820                     "CPU to profile on"),
1821         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
1822                     "dump raw trace in ASCII"),
1823         OPT_END()
1824 };
1825
1826 static const char * const replay_usage[] = {
1827         "perf sched replay [<options>]",
1828         NULL
1829 };
1830
1831 static const struct option replay_options[] = {
1832         OPT_INTEGER('r', "repeat", &replay_repeat,
1833                     "repeat the workload replay N times (-1: infinite)"),
1834         OPT_BOOLEAN('v', "verbose", &verbose,
1835                     "be more verbose (show symbol address, etc)"),
1836         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
1837                     "dump raw trace in ASCII"),
1838         OPT_END()
1839 };
1840
1841 static void setup_sorting(void)
1842 {
1843         char *tmp, *tok, *str = strdup(sort_order);
1844
1845         for (tok = strtok_r(str, ", ", &tmp);
1846                         tok; tok = strtok_r(NULL, ", ", &tmp)) {
1847                 if (sort_dimension__add(tok, &sort_list) < 0) {
1848                         error("Unknown --sort key: `%s'", tok);
1849                         usage_with_options(latency_usage, latency_options);
1850                 }
1851         }
1852
1853         free(str);
1854
1855         sort_dimension__add("pid", &cmp_pid);
1856 }
1857
1858 static const char *record_args[] = {
1859         "record",
1860         "-a",
1861         "-R",
1862         "-M",
1863         "-f",
1864         "-m", "1024",
1865         "-c", "1",
1866         "-e", "sched:sched_switch:r",
1867         "-e", "sched:sched_stat_wait:r",
1868         "-e", "sched:sched_stat_sleep:r",
1869         "-e", "sched:sched_stat_iowait:r",
1870         "-e", "sched:sched_stat_runtime:r",
1871         "-e", "sched:sched_process_exit:r",
1872         "-e", "sched:sched_process_fork:r",
1873         "-e", "sched:sched_wakeup:r",
1874         "-e", "sched:sched_migrate_task:r",
1875 };
1876
1877 static int __cmd_record(int argc, const char **argv)
1878 {
1879         unsigned int rec_argc, i, j;
1880         const char **rec_argv;
1881
1882         rec_argc = ARRAY_SIZE(record_args) + argc - 1;
1883         rec_argv = calloc(rec_argc + 1, sizeof(char *));
1884
1885         for (i = 0; i < ARRAY_SIZE(record_args); i++)
1886                 rec_argv[i] = strdup(record_args[i]);
1887
1888         for (j = 1; j < (unsigned int)argc; j++, i++)
1889                 rec_argv[i] = argv[j];
1890
1891         BUG_ON(i != rec_argc);
1892
1893         return cmd_record(i, rec_argv, NULL);
1894 }
1895
1896 int cmd_sched(int argc, const char **argv, const char *prefix __used)
1897 {
1898         argc = parse_options(argc, argv, sched_options, sched_usage,
1899                              PARSE_OPT_STOP_AT_NON_OPTION);
1900         if (!argc)
1901                 usage_with_options(sched_usage, sched_options);
1902
1903         /*
1904          * Aliased to 'perf trace' for now:
1905          */
1906         if (!strcmp(argv[0], "trace"))
1907                 return cmd_trace(argc, argv, prefix);
1908
1909         symbol__init(0);
1910         if (!strncmp(argv[0], "rec", 3)) {
1911                 return __cmd_record(argc, argv);
1912         } else if (!strncmp(argv[0], "lat", 3)) {
1913                 trace_handler = &lat_ops;
1914                 if (argc > 1) {
1915                         argc = parse_options(argc, argv, latency_options, latency_usage, 0);
1916                         if (argc)
1917                                 usage_with_options(latency_usage, latency_options);
1918                 }
1919                 setup_sorting();
1920                 __cmd_lat();
1921         } else if (!strcmp(argv[0], "map")) {
1922                 trace_handler = &map_ops;
1923                 setup_sorting();
1924                 __cmd_map();
1925         } else if (!strncmp(argv[0], "rep", 3)) {
1926                 trace_handler = &replay_ops;
1927                 if (argc) {
1928                         argc = parse_options(argc, argv, replay_options, replay_usage, 0);
1929                         if (argc)
1930                                 usage_with_options(replay_usage, replay_options);
1931                 }
1932                 __cmd_replay();
1933         } else {
1934                 usage_with_options(sched_usage, sched_options);
1935         }
1936
1937         return 0;
1938 }