tools/perf/builtin-sched.c

   1 // SPDX-License-Identifier: GPL-2.0
   2 #include "builtin.h"
   3 #include "perf.h"
   4 #include "perf-sys.h"
   5
   6 #include "util/cpumap.h"
   7 #include "util/evlist.h"
   8 #include "util/evsel.h"
   9 #include "util/evsel_fprintf.h"
  10 #include "util/symbol.h"
  11 #include "util/thread.h"
  12 #include "util/header.h"
  13 #include "util/session.h"
  14 #include "util/tool.h"
  15 #include "util/cloexec.h"
  16 #include "util/thread_map.h"
  17 #include "util/color.h"
  18 #include "util/stat.h"
  19 #include "util/string2.h"
  20 #include "util/callchain.h"
  21 #include "util/time-utils.h"
  22
  23 #include <subcmd/pager.h>
  24 #include <subcmd/parse-options.h>
  25 #include "util/trace-event.h"
  26
  27 #include "util/debug.h"
  28 #include "util/event.h"
  29
  30 #include <linux/kernel.h>
  31 #include <linux/log2.h>
  32 #include <linux/zalloc.h>
  33 #include <sys/prctl.h>
  34 #include <sys/resource.h>
  35 #include <inttypes.h>
  36
  37 #include <errno.h>
  38 #include <semaphore.h>
  39 #include <pthread.h>
  40 #include <math.h>
  41 #include <api/fs/fs.h>
  42 #include <perf/cpumap.h>
  43 #include <linux/time64.h>
  44 #include <linux/err.h>
  45
  46 #include <linux/ctype.h>
  47
  48 #define PR_SET_NAME             15               /* Set process name */
  49 #define MAX_CPUS                4096
  50 #define COMM_LEN                20
  51 #define SYM_LEN                 129
  52 #define MAX_PID                 1024000
  53
  54 static const char *cpu_list;
  55 static DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS);
  56
  57 struct sched_atom;
  58
  59 struct task_desc {
  60         unsigned long           nr;
  61         unsigned long           pid;
  62         char                    comm[COMM_LEN];
  63
  64         unsigned long           nr_events;
  65         unsigned long           curr_event;
  66         struct sched_atom       **atoms;
  67
  68         pthread_t               thread;
  69         sem_t                   sleep_sem;
  70
  71         sem_t                   ready_for_work;
  72         sem_t                   work_done_sem;
  73
  74         u64                     cpu_usage;
  75 };
  76
  77 enum sched_event_type {
  78         SCHED_EVENT_RUN,
  79         SCHED_EVENT_SLEEP,
  80         SCHED_EVENT_WAKEUP,
  81         SCHED_EVENT_MIGRATION,
  82 };
  83
  84 struct sched_atom {
  85         enum sched_event_type   type;
  86         int                     specific_wait;
  87         u64                     timestamp;
  88         u64                     duration;
  89         unsigned long           nr;
  90         sem_t                   *wait_sem;
  91         struct task_desc        *wakee;
  92 };
  93
  94 #define TASK_STATE_TO_CHAR_STR "RSDTtZXxKWP"
  95
  96 /* task state bitmask, copied from include/linux/sched.h */
  97 #define TASK_RUNNING            0
  98 #define TASK_INTERRUPTIBLE      1
  99 #define TASK_UNINTERRUPTIBLE    2
 100 #define __TASK_STOPPED          4
 101 #define __TASK_TRACED           8
 102 /* in tsk->exit_state */
 103 #define EXIT_DEAD               16
 104 #define EXIT_ZOMBIE             32
 105 #define EXIT_TRACE              (EXIT_ZOMBIE | EXIT_DEAD)
 106 /* in tsk->state again */
 107 #define TASK_DEAD               64
 108 #define TASK_WAKEKILL           128
 109 #define TASK_WAKING             256
 110 #define TASK_PARKED             512
 111
 112 enum thread_state {
 113         THREAD_SLEEPING = 0,
 114         THREAD_WAIT_CPU,
 115         THREAD_SCHED_IN,
 116         THREAD_IGNORE
 117 };
 118
 119 struct work_atom {
 120         struct list_head        list;
 121         enum thread_state       state;
 122         u64                     sched_out_time;
 123         u64                     wake_up_time;
 124         u64                     sched_in_time;
 125         u64                     runtime;
 126 };
 127
 128 struct work_atoms {
 129         struct list_head        work_list;
 130         struct thread           *thread;
 131         struct rb_node          node;
 132         u64                     max_lat;
 133         u64                     max_lat_start;
 134         u64                     max_lat_end;
 135         u64                     total_lat;
 136         u64                     nb_atoms;
 137         u64                     total_runtime;
 138         int                     num_merged;
 139 };
 140
 141 typedef int (*sort_fn_t)(struct work_atoms *, struct work_atoms *);
 142
 143 struct perf_sched;
 144
 145 struct trace_sched_handler {
 146         int (*switch_event)(struct perf_sched *sched, struct evsel *evsel,
 147                             struct perf_sample *sample, struct machine *machine);
 148
 149         int (*runtime_event)(struct perf_sched *sched, struct evsel *evsel,
 150                              struct perf_sample *sample, struct machine *machine);
 151
 152         int (*wakeup_event)(struct perf_sched *sched, struct evsel *evsel,
 153                             struct perf_sample *sample, struct machine *machine);
 154
 155         /* PERF_RECORD_FORK event, not sched_process_fork tracepoint */
 156         int (*fork_event)(struct perf_sched *sched, union perf_event *event,
 157                           struct machine *machine);
 158
 159         int (*migrate_task_event)(struct perf_sched *sched,
 160                                   struct evsel *evsel,
 161                                   struct perf_sample *sample,
 162                                   struct machine *machine);
 163 };
 164
 165 #define COLOR_PIDS PERF_COLOR_BLUE
 166 #define COLOR_CPUS PERF_COLOR_BG_RED
 167
 168 struct perf_sched_map {
 169         DECLARE_BITMAP(comp_cpus_mask, MAX_CPUS);
 170         struct perf_cpu         *comp_cpus;
 171         bool                     comp;
 172         struct perf_thread_map *color_pids;
 173         const char              *color_pids_str;
 174         struct perf_cpu_map     *color_cpus;
 175         const char              *color_cpus_str;
 176         struct perf_cpu_map     *cpus;
 177         const char              *cpus_str;
 178 };
 179
 180 struct perf_sched {
 181         struct perf_tool tool;
 182         const char       *sort_order;
 183         unsigned long    nr_tasks;
 184         struct task_desc **pid_to_task;
 185         struct task_desc **tasks;
 186         const struct trace_sched_handler *tp_handler;
 187         pthread_mutex_t  start_work_mutex;
 188         pthread_mutex_t  work_done_wait_mutex;
 189         int              profile_cpu;
 190 /*
 191  * Track the current task - that way we can know whether there's any
 192  * weird events, such as a task being switched away that is not current.
 193  */
 194         struct perf_cpu  max_cpu;
 195         u32              curr_pid[MAX_CPUS];
 196         struct thread    *curr_thread[MAX_CPUS];
 197         char             next_shortname1;
 198         char             next_shortname2;
 199         unsigned int     replay_repeat;
 200         unsigned long    nr_run_events;
 201         unsigned long    nr_sleep_events;
 202         unsigned long    nr_wakeup_events;
 203         unsigned long    nr_sleep_corrections;
 204         unsigned long    nr_run_events_optimized;
 205         unsigned long    targetless_wakeups;
 206         unsigned long    multitarget_wakeups;
 207         unsigned long    nr_runs;
 208         unsigned long    nr_timestamps;
 209         unsigned long    nr_unordered_timestamps;
 210         unsigned long    nr_context_switch_bugs;
 211         unsigned long    nr_events;
 212         unsigned long    nr_lost_chunks;
 213         unsigned long    nr_lost_events;
 214         u64              run_measurement_overhead;
 215         u64              sleep_measurement_overhead;
 216         u64              start_time;
 217         u64              cpu_usage;
 218         u64              runavg_cpu_usage;
 219         u64              parent_cpu_usage;
 220         u64              runavg_parent_cpu_usage;
 221         u64              sum_runtime;
 222         u64              sum_fluct;
 223         u64              run_avg;
 224         u64              all_runtime;
 225         u64              all_count;
 226         u64              cpu_last_switched[MAX_CPUS];
 227         struct rb_root_cached atom_root, sorted_atom_root, merged_atom_root;
 228         struct list_head sort_list, cmp_pid;
 229         bool force;
 230         bool skip_merge;
 231         struct perf_sched_map map;
 232
 233         /* options for timehist command */
 234         bool            summary;
 235         bool            summary_only;
 236         bool            idle_hist;
 237         bool            show_callchain;
 238         unsigned int    max_stack;
 239         bool            show_cpu_visual;
 240         bool            show_wakeups;
 241         bool            show_next;
 242         bool            show_migrations;
 243         bool            show_state;
 244         u64             skipped_samples;
 245         const char      *time_str;
 246         struct perf_time_interval ptime;
 247         struct perf_time_interval hist_time;
 248 };
 249
 250 /* per thread run time data */
 251 struct thread_runtime {
 252         u64 last_time;      /* time of previous sched in/out event */
 253         u64 dt_run;         /* run time */
 254         u64 dt_sleep;       /* time between CPU access by sleep (off cpu) */
 255         u64 dt_iowait;      /* time between CPU access by iowait (off cpu) */
 256         u64 dt_preempt;     /* time between CPU access by preempt (off cpu) */
 257         u64 dt_delay;       /* time between wakeup and sched-in */
 258         u64 ready_to_run;   /* time of wakeup */
 259
 260         struct stats run_stats;
 261         u64 total_run_time;
 262         u64 total_sleep_time;
 263         u64 total_iowait_time;
 264         u64 total_preempt_time;
 265         u64 total_delay_time;
 266
 267         int last_state;
 268
 269         char shortname[3];
 270         bool comm_changed;
 271
 272         u64 migrations;
 273 };
 274
 275 /* per event run time data */
 276 struct evsel_runtime {
 277         u64 *last_time; /* time this event was last seen per cpu */
 278         u32 ncpu;       /* highest cpu slot allocated */
 279 };
 280
 281 /* per cpu idle time data */
 282 struct idle_thread_runtime {
 283         struct thread_runtime   tr;
 284         struct thread           *last_thread;
 285         struct rb_root_cached   sorted_root;
 286         struct callchain_root   callchain;
 287         struct callchain_cursor cursor;
 288 };
 289
 290 /* track idle times per cpu */
 291 static struct thread **idle_threads;
 292 static int idle_max_cpu;
 293 static char idle_comm[] = "<idle>";
 294
 295 static u64 get_nsecs(void)
 296 {
 297         struct timespec ts;
 298
 299         clock_gettime(CLOCK_MONOTONIC, &ts);
 300
 301         return ts.tv_sec * NSEC_PER_SEC + ts.tv_nsec;
 302 }
 303
 304 static void burn_nsecs(struct perf_sched *sched, u64 nsecs)
 305 {
 306         u64 T0 = get_nsecs(), T1;
 307
 308         do {
 309                 T1 = get_nsecs();
 310         } while (T1 + sched->run_measurement_overhead < T0 + nsecs);
 311 }
 312
 313 static void sleep_nsecs(u64 nsecs)
 314 {
 315         struct timespec ts;
 316
 317         ts.tv_nsec = nsecs % 999999999;
 318         ts.tv_sec = nsecs / 999999999;
 319
 320         nanosleep(&ts, NULL);
 321 }
 322
 323 static void calibrate_run_measurement_overhead(struct perf_sched *sched)
 324 {
 325         u64 T0, T1, delta, min_delta = NSEC_PER_SEC;
 326         int i;
 327
 328         for (i = 0; i < 10; i++) {
 329                 T0 = get_nsecs();
 330                 burn_nsecs(sched, 0);
 331                 T1 = get_nsecs();
 332                 delta = T1-T0;
 333                 min_delta = min(min_delta, delta);
 334         }
 335         sched->run_measurement_overhead = min_delta;
 336
 337         printf("run measurement overhead: %" PRIu64 " nsecs\n", min_delta);
 338 }
 339
 340 static void calibrate_sleep_measurement_overhead(struct perf_sched *sched)
 341 {
 342         u64 T0, T1, delta, min_delta = NSEC_PER_SEC;
 343         int i;
 344
 345         for (i = 0; i < 10; i++) {
 346                 T0 = get_nsecs();
 347                 sleep_nsecs(10000);
 348                 T1 = get_nsecs();
 349                 delta = T1-T0;
 350                 min_delta = min(min_delta, delta);
 351         }
 352         min_delta -= 10000;
 353         sched->sleep_measurement_overhead = min_delta;
 354
 355         printf("sleep measurement overhead: %" PRIu64 " nsecs\n", min_delta);
 356 }
 357
 358 static struct sched_atom *
 359 get_new_event(struct task_desc *task, u64 timestamp)
 360 {
 361         struct sched_atom *event = zalloc(sizeof(*event));
 362         unsigned long idx = task->nr_events;
 363         size_t size;
 364
 365         event->timestamp = timestamp;
 366         event->nr = idx;
 367
 368         task->nr_events++;
 369         size = sizeof(struct sched_atom *) * task->nr_events;
 370         task->atoms = realloc(task->atoms, size);
 371         BUG_ON(!task->atoms);
 372
 373         task->atoms[idx] = event;
 374
 375         return event;
 376 }
 377
 378 static struct sched_atom *last_event(struct task_desc *task)
 379 {
 380         if (!task->nr_events)
 381                 return NULL;
 382
 383         return task->atoms[task->nr_events - 1];
 384 }
 385
 386 static void add_sched_event_run(struct perf_sched *sched, struct task_desc *task,
 387                                 u64 timestamp, u64 duration)
 388 {
 389         struct sched_atom *event, *curr_event = last_event(task);
 390
 391         /*
 392          * optimize an existing RUN event by merging this one
 393          * to it:
 394          */
 395         if (curr_event && curr_event->type == SCHED_EVENT_RUN) {
 396                 sched->nr_run_events_optimized++;
 397                 curr_event->duration += duration;
 398                 return;
 399         }
 400
 401         event = get_new_event(task, timestamp);
 402
 403         event->type = SCHED_EVENT_RUN;
 404         event->duration = duration;
 405
 406         sched->nr_run_events++;
 407 }
 408
 409 static void add_sched_event_wakeup(struct perf_sched *sched, struct task_desc *task,
 410                                    u64 timestamp, struct task_desc *wakee)
 411 {
 412         struct sched_atom *event, *wakee_event;
 413
 414         event = get_new_event(task, timestamp);
 415         event->type = SCHED_EVENT_WAKEUP;
 416         event->wakee = wakee;
 417
 418         wakee_event = last_event(wakee);
 419         if (!wakee_event || wakee_event->type != SCHED_EVENT_SLEEP) {
 420                 sched->targetless_wakeups++;
 421                 return;
 422         }
 423         if (wakee_event->wait_sem) {
 424                 sched->multitarget_wakeups++;
 425                 return;
 426         }
 427
 428         wakee_event->wait_sem = zalloc(sizeof(*wakee_event->wait_sem));
 429         sem_init(wakee_event->wait_sem, 0, 0);
 430         wakee_event->specific_wait = 1;
 431         event->wait_sem = wakee_event->wait_sem;
 432
 433         sched->nr_wakeup_events++;
 434 }
 435
 436 static void add_sched_event_sleep(struct perf_sched *sched, struct task_desc *task,
 437                                   u64 timestamp, u64 task_state __maybe_unused)
 438 {
 439         struct sched_atom *event = get_new_event(task, timestamp);
 440
 441         event->type = SCHED_EVENT_SLEEP;
 442
 443         sched->nr_sleep_events++;
 444 }
 445
 446 static struct task_desc *register_pid(struct perf_sched *sched,
 447                                       unsigned long pid, const char *comm)
 448 {
 449         struct task_desc *task;
 450         static int pid_max;
 451
 452         if (sched->pid_to_task == NULL) {
 453                 if (sysctl__read_int("kernel/pid_max", &pid_max) < 0)
 454                         pid_max = MAX_PID;
 455                 BUG_ON((sched->pid_to_task = calloc(pid_max, sizeof(struct task_desc *))) == NULL);
 456         }
 457         if (pid >= (unsigned long)pid_max) {
 458                 BUG_ON((sched->pid_to_task = realloc(sched->pid_to_task, (pid + 1) *
 459                         sizeof(struct task_desc *))) == NULL);
 460                 while (pid >= (unsigned long)pid_max)
 461                         sched->pid_to_task[pid_max++] = NULL;
 462         }
 463
 464         task = sched->pid_to_task[pid];
 465
 466         if (task)
 467                 return task;
 468
 469         task = zalloc(sizeof(*task));
 470         task->pid = pid;
 471         task->nr = sched->nr_tasks;
 472         strcpy(task->comm, comm);
 473         /*
 474          * every task starts in sleeping state - this gets ignored
 475          * if there's no wakeup pointing to this sleep state:
 476          */
 477         add_sched_event_sleep(sched, task, 0, 0);
 478
 479         sched->pid_to_task[pid] = task;
 480         sched->nr_tasks++;
 481         sched->tasks = realloc(sched->tasks, sched->nr_tasks * sizeof(struct task_desc *));
 482         BUG_ON(!sched->tasks);
 483         sched->tasks[task->nr] = task;
 484
 485         if (verbose > 0)
 486                 printf("registered task #%ld, PID %ld (%s)\n", sched->nr_tasks, pid, comm);
 487
 488         return task;
 489 }
 490
 491
 492 static void print_task_traces(struct perf_sched *sched)
 493 {
 494         struct task_desc *task;
 495         unsigned long i;
 496
 497         for (i = 0; i < sched->nr_tasks; i++) {
 498                 task = sched->tasks[i];
 499                 printf("task %6ld (%20s:%10ld), nr_events: %ld\n",
 500                         task->nr, task->comm, task->pid, task->nr_events);
 501         }
 502 }
 503
 504 static void add_cross_task_wakeups(struct perf_sched *sched)
 505 {
 506         struct task_desc *task1, *task2;
 507         unsigned long i, j;
 508
 509         for (i = 0; i < sched->nr_tasks; i++) {
 510                 task1 = sched->tasks[i];
 511                 j = i + 1;
 512                 if (j == sched->nr_tasks)
 513                         j = 0;
 514                 task2 = sched->tasks[j];
 515                 add_sched_event_wakeup(sched, task1, 0, task2);
 516         }
 517 }
 518
 519 static void perf_sched__process_event(struct perf_sched *sched,
 520                                       struct sched_atom *atom)
 521 {
 522         int ret = 0;
 523
 524         switch (atom->type) {
 525                 case SCHED_EVENT_RUN:
 526                         burn_nsecs(sched, atom->duration);
 527                         break;
 528                 case SCHED_EVENT_SLEEP:
 529                         if (atom->wait_sem)
 530                                 ret = sem_wait(atom->wait_sem);
 531                         BUG_ON(ret);
 532                         break;
 533                 case SCHED_EVENT_WAKEUP:
 534                         if (atom->wait_sem)
 535                                 ret = sem_post(atom->wait_sem);
 536                         BUG_ON(ret);
 537                         break;
 538                 case SCHED_EVENT_MIGRATION:
 539                         break;
 540                 default:
 541                         BUG_ON(1);
 542         }
 543 }
 544
 545 static u64 get_cpu_usage_nsec_parent(void)
 546 {
 547         struct rusage ru;
 548         u64 sum;
 549         int err;
 550
 551         err = getrusage(RUSAGE_SELF, &ru);
 552         BUG_ON(err);
 553
 554         sum =  ru.ru_utime.tv_sec * NSEC_PER_SEC + ru.ru_utime.tv_usec * NSEC_PER_USEC;
 555         sum += ru.ru_stime.tv_sec * NSEC_PER_SEC + ru.ru_stime.tv_usec * NSEC_PER_USEC;
 556
 557         return sum;
 558 }
 559
 560 static int self_open_counters(struct perf_sched *sched, unsigned long cur_task)
 561 {
 562         struct perf_event_attr attr;
 563         char sbuf[STRERR_BUFSIZE], info[STRERR_BUFSIZE];
 564         int fd;
 565         struct rlimit limit;
 566         bool need_privilege = false;
 567
 568         memset(&attr, 0, sizeof(attr));
 569
 570         attr.type = PERF_TYPE_SOFTWARE;
 571         attr.config = PERF_COUNT_SW_TASK_CLOCK;
 572
 573 force_again:
 574         fd = sys_perf_event_open(&attr, 0, -1, -1,
 575                                  perf_event_open_cloexec_flag());
 576
 577         if (fd < 0) {
 578                 if (errno == EMFILE) {
 579                         if (sched->force) {
 580                                 BUG_ON(getrlimit(RLIMIT_NOFILE, &limit) == -1);
 581                                 limit.rlim_cur += sched->nr_tasks - cur_task;
 582                                 if (limit.rlim_cur > limit.rlim_max) {
 583                                         limit.rlim_max = limit.rlim_cur;
 584                                         need_privilege = true;
 585                                 }
 586                                 if (setrlimit(RLIMIT_NOFILE, &limit) == -1) {
 587                                         if (need_privilege && errno == EPERM)
 588                                                 strcpy(info, "Need privilege\n");
 589                                 } else
 590                                         goto force_again;
 591                         } else
 592                                 strcpy(info, "Have a try with -f option\n");
 593                 }
 594                 pr_err("Error: sys_perf_event_open() syscall returned "
 595                        "with %d (%s)\n%s", fd,
 596                        str_error_r(errno, sbuf, sizeof(sbuf)), info);
 597                 exit(EXIT_FAILURE);
 598         }
 599         return fd;
 600 }
 601
 602 static u64 get_cpu_usage_nsec_self(int fd)
 603 {
 604         u64 runtime;
 605         int ret;
 606
 607         ret = read(fd, &runtime, sizeof(runtime));
 608         BUG_ON(ret != sizeof(runtime));
 609
 610         return runtime;
 611 }
 612
 613 struct sched_thread_parms {
 614         struct task_desc  *task;
 615         struct perf_sched *sched;
 616         int fd;
 617 };
 618
 619 static void *thread_func(void *ctx)
 620 {
 621         struct sched_thread_parms *parms = ctx;
 622         struct task_desc *this_task = parms->task;
 623         struct perf_sched *sched = parms->sched;
 624         u64 cpu_usage_0, cpu_usage_1;
 625         unsigned long i, ret;
 626         char comm2[22];
 627         int fd = parms->fd;
 628
 629         zfree(&parms);
 630
 631         sprintf(comm2, ":%s", this_task->comm);
 632         prctl(PR_SET_NAME, comm2);
 633         if (fd < 0)
 634                 return NULL;
 635 again:
 636         ret = sem_post(&this_task->ready_for_work);
 637         BUG_ON(ret);
 638         ret = pthread_mutex_lock(&sched->start_work_mutex);
 639         BUG_ON(ret);
 640         ret = pthread_mutex_unlock(&sched->start_work_mutex);
 641         BUG_ON(ret);
 642
 643         cpu_usage_0 = get_cpu_usage_nsec_self(fd);
 644
 645         for (i = 0; i < this_task->nr_events; i++) {
 646                 this_task->curr_event = i;
 647                 perf_sched__process_event(sched, this_task->atoms[i]);
 648         }
 649
 650         cpu_usage_1 = get_cpu_usage_nsec_self(fd);
 651         this_task->cpu_usage = cpu_usage_1 - cpu_usage_0;
 652         ret = sem_post(&this_task->work_done_sem);
 653         BUG_ON(ret);
 654
 655         ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
 656         BUG_ON(ret);
 657         ret = pthread_mutex_unlock(&sched->work_done_wait_mutex);
 658         BUG_ON(ret);
 659
 660         goto again;
 661 }
 662
 663 static void create_tasks(struct perf_sched *sched)
 664 {
 665         struct task_desc *task;
 666         pthread_attr_t attr;
 667         unsigned long i;
 668         int err;
 669
 670         err = pthread_attr_init(&attr);
 671         BUG_ON(err);
 672         err = pthread_attr_setstacksize(&attr,
 673                         (size_t) max(16 * 1024, (int)PTHREAD_STACK_MIN));
 674         BUG_ON(err);
 675         err = pthread_mutex_lock(&sched->start_work_mutex);
 676         BUG_ON(err);
 677         err = pthread_mutex_lock(&sched->work_done_wait_mutex);
 678         BUG_ON(err);
 679         for (i = 0; i < sched->nr_tasks; i++) {
 680                 struct sched_thread_parms *parms = malloc(sizeof(*parms));
 681                 BUG_ON(parms == NULL);
 682                 parms->task = task = sched->tasks[i];
 683                 parms->sched = sched;
 684                 parms->fd = self_open_counters(sched, i);
 685                 sem_init(&task->sleep_sem, 0, 0);
 686                 sem_init(&task->ready_for_work, 0, 0);
 687                 sem_init(&task->work_done_sem, 0, 0);
 688                 task->curr_event = 0;
 689                 err = pthread_create(&task->thread, &attr, thread_func, parms);
 690                 BUG_ON(err);
 691         }
 692 }
 693
 694 static void wait_for_tasks(struct perf_sched *sched)
 695 {
 696         u64 cpu_usage_0, cpu_usage_1;
 697         struct task_desc *task;
 698         unsigned long i, ret;
 699
 700         sched->start_time = get_nsecs();
 701         sched->cpu_usage = 0;
 702         pthread_mutex_unlock(&sched->work_done_wait_mutex);
 703
 704         for (i = 0; i < sched->nr_tasks; i++) {
 705                 task = sched->tasks[i];
 706                 ret = sem_wait(&task->ready_for_work);
 707                 BUG_ON(ret);
 708                 sem_init(&task->ready_for_work, 0, 0);
 709         }
 710         ret = pthread_mutex_lock(&sched->work_done_wait_mutex);
 711         BUG_ON(ret);
 712
 713         cpu_usage_0 = get_cpu_usage_nsec_parent();
 714
 715         pthread_mutex_unlock(&sched->start_work_mutex);
 716
 717         for (i = 0; i < sched->nr_tasks; i++) {
 718                 task = sched->tasks[i];
 719                 ret = sem_wait(&task->work_done_sem);
 720                 BUG_ON(ret);
 721                 sem_init(&task->work_done_sem, 0, 0);
 722                 sched->cpu_usage += task->cpu_usage;
 723                 task->cpu_usage = 0;
 724         }
 725
 726         cpu_usage_1 = get_cpu_usage_nsec_parent();
 727         if (!sched->runavg_cpu_usage)
 728                 sched->runavg_cpu_usage = sched->cpu_usage;
 729         sched->runavg_cpu_usage = (sched->runavg_cpu_usage * (sched->replay_repeat - 1) + sched->cpu_usage) / sched->replay_repeat;
 730
 731         sched->parent_cpu_usage = cpu_usage_1 - cpu_usage_0;
 732         if (!sched->runavg_parent_cpu_usage)
 733                 sched->runavg_parent_cpu_usage = sched->parent_cpu_usage;
 734         sched->runavg_parent_cpu_usage = (sched->runavg_parent_cpu_usage * (sched->replay_repeat - 1) +
 735                                          sched->parent_cpu_usage)/sched->replay_repeat;
 736
 737         ret = pthread_mutex_lock(&sched->start_work_mutex);
 738         BUG_ON(ret);
 739
 740         for (i = 0; i < sched->nr_tasks; i++) {
 741                 task = sched->tasks[i];
 742                 sem_init(&task->sleep_sem, 0, 0);
 743                 task->curr_event = 0;
 744         }
 745 }
 746
 747 static void run_one_test(struct perf_sched *sched)
 748 {
 749         u64 T0, T1, delta, avg_delta, fluct;
 750
 751         T0 = get_nsecs();
 752         wait_for_tasks(sched);
 753         T1 = get_nsecs();
 754
 755         delta = T1 - T0;
 756         sched->sum_runtime += delta;
 757         sched->nr_runs++;
 758
 759         avg_delta = sched->sum_runtime / sched->nr_runs;
 760         if (delta < avg_delta)
 761                 fluct = avg_delta - delta;
 762         else
 763                 fluct = delta - avg_delta;
 764         sched->sum_fluct += fluct;
 765         if (!sched->run_avg)
 766                 sched->run_avg = delta;
 767         sched->run_avg = (sched->run_avg * (sched->replay_repeat - 1) + delta) / sched->replay_repeat;
 768
 769         printf("#%-3ld: %0.3f, ", sched->nr_runs, (double)delta / NSEC_PER_MSEC);
 770
 771         printf("ravg: %0.2f, ", (double)sched->run_avg / NSEC_PER_MSEC);
 772
 773         printf("cpu: %0.2f / %0.2f",
 774                 (double)sched->cpu_usage / NSEC_PER_MSEC, (double)sched->runavg_cpu_usage / NSEC_PER_MSEC);
 775
 776 #if 0
 777         /*
 778          * rusage statistics done by the parent, these are less
 779          * accurate than the sched->sum_exec_runtime based statistics:
 780          */
 781         printf(" [%0.2f / %0.2f]",
 782                 (double)sched->parent_cpu_usage / NSEC_PER_MSEC,
 783                 (double)sched->runavg_parent_cpu_usage / NSEC_PER_MSEC);
 784 #endif
 785
 786         printf("\n");
 787
 788         if (sched->nr_sleep_corrections)
 789                 printf(" (%ld sleep corrections)\n", sched->nr_sleep_corrections);
 790         sched->nr_sleep_corrections = 0;
 791 }
 792
 793 static void test_calibrations(struct perf_sched *sched)
 794 {
 795         u64 T0, T1;
 796
 797         T0 = get_nsecs();
 798         burn_nsecs(sched, NSEC_PER_MSEC);
 799         T1 = get_nsecs();
 800
 801         printf("the run test took %" PRIu64 " nsecs\n", T1 - T0);
 802
 803         T0 = get_nsecs();
 804         sleep_nsecs(NSEC_PER_MSEC);
 805         T1 = get_nsecs();
 806
 807         printf("the sleep test took %" PRIu64 " nsecs\n", T1 - T0);
 808 }
 809
 810 static int
 811 replay_wakeup_event(struct perf_sched *sched,
 812                     struct evsel *evsel, struct perf_sample *sample,
 813                     struct machine *machine __maybe_unused)
 814 {
 815         const char *comm = evsel__strval(evsel, sample, "comm");
 816         const u32 pid    = evsel__intval(evsel, sample, "pid");
 817         struct task_desc *waker, *wakee;
 818
 819         if (verbose > 0) {
 820                 printf("sched_wakeup event %p\n", evsel);
 821
 822                 printf(" ... pid %d woke up %s/%d\n", sample->tid, comm, pid);
 823         }
 824
 825         waker = register_pid(sched, sample->tid, "<unknown>");
 826         wakee = register_pid(sched, pid, comm);
 827
 828         add_sched_event_wakeup(sched, waker, sample->time, wakee);
 829         return 0;
 830 }
 831
 832 static int replay_switch_event(struct perf_sched *sched,
 833                                struct evsel *evsel,
 834                                struct perf_sample *sample,
 835                                struct machine *machine __maybe_unused)
 836 {
 837         const char *prev_comm  = evsel__strval(evsel, sample, "prev_comm"),
 838                    *next_comm  = evsel__strval(evsel, sample, "next_comm");
 839         const u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
 840                   next_pid = evsel__intval(evsel, sample, "next_pid");
 841         const u64 prev_state = evsel__intval(evsel, sample, "prev_state");
 842         struct task_desc *prev, __maybe_unused *next;
 843         u64 timestamp0, timestamp = sample->time;
 844         int cpu = sample->cpu;
 845         s64 delta;
 846
 847         if (verbose > 0)
 848                 printf("sched_switch event %p\n", evsel);
 849
 850         if (cpu >= MAX_CPUS || cpu < 0)
 851                 return 0;
 852
 853         timestamp0 = sched->cpu_last_switched[cpu];
 854         if (timestamp0)
 855                 delta = timestamp - timestamp0;
 856         else
 857                 delta = 0;
 858
 859         if (delta < 0) {
 860                 pr_err("hm, delta: %" PRIu64 " < 0 ?\n", delta);
 861                 return -1;
 862         }
 863
 864         pr_debug(" ... switch from %s/%d to %s/%d [ran %" PRIu64 " nsecs]\n",
 865                  prev_comm, prev_pid, next_comm, next_pid, delta);
 866
 867         prev = register_pid(sched, prev_pid, prev_comm);
 868         next = register_pid(sched, next_pid, next_comm);
 869
 870         sched->cpu_last_switched[cpu] = timestamp;
 871
 872         add_sched_event_run(sched, prev, timestamp, delta);
 873         add_sched_event_sleep(sched, prev, timestamp, prev_state);
 874
 875         return 0;
 876 }
 877
 878 static int replay_fork_event(struct perf_sched *sched,
 879                              union perf_event *event,
 880                              struct machine *machine)
 881 {
 882         struct thread *child, *parent;
 883
 884         child = machine__findnew_thread(machine, event->fork.pid,
 885                                         event->fork.tid);
 886         parent = machine__findnew_thread(machine, event->fork.ppid,
 887                                          event->fork.ptid);
 888
 889         if (child == NULL || parent == NULL) {
 890                 pr_debug("thread does not exist on fork event: child %p, parent %p\n",
 891                                  child, parent);
 892                 goto out_put;
 893         }
 894
 895         if (verbose > 0) {
 896                 printf("fork event\n");
 897                 printf("... parent: %s/%d\n", thread__comm_str(parent), parent->tid);
 898                 printf("...  child: %s/%d\n", thread__comm_str(child), child->tid);
 899         }
 900
 901         register_pid(sched, parent->tid, thread__comm_str(parent));
 902         register_pid(sched, child->tid, thread__comm_str(child));
 903 out_put:
 904         thread__put(child);
 905         thread__put(parent);
 906         return 0;
 907 }
 908
 909 struct sort_dimension {
 910         const char              *name;
 911         sort_fn_t               cmp;
 912         struct list_head        list;
 913 };
 914
 915 /*
 916  * handle runtime stats saved per thread
 917  */
 918 static struct thread_runtime *thread__init_runtime(struct thread *thread)
 919 {
 920         struct thread_runtime *r;
 921
 922         r = zalloc(sizeof(struct thread_runtime));
 923         if (!r)
 924                 return NULL;
 925
 926         init_stats(&r->run_stats);
 927         thread__set_priv(thread, r);
 928
 929         return r;
 930 }
 931
 932 static struct thread_runtime *thread__get_runtime(struct thread *thread)
 933 {
 934         struct thread_runtime *tr;
 935
 936         tr = thread__priv(thread);
 937         if (tr == NULL) {
 938                 tr = thread__init_runtime(thread);
 939                 if (tr == NULL)
 940                         pr_debug("Failed to malloc memory for runtime data.\n");
 941         }
 942
 943         return tr;
 944 }
 945
 946 static int
 947 thread_lat_cmp(struct list_head *list, struct work_atoms *l, struct work_atoms *r)
 948 {
 949         struct sort_dimension *sort;
 950         int ret = 0;
 951
 952         BUG_ON(list_empty(list));
 953
 954         list_for_each_entry(sort, list, list) {
 955                 ret = sort->cmp(l, r);
 956                 if (ret)
 957                         return ret;
 958         }
 959
 960         return ret;
 961 }
 962
 963 static struct work_atoms *
 964 thread_atoms_search(struct rb_root_cached *root, struct thread *thread,
 965                          struct list_head *sort_list)
 966 {
 967         struct rb_node *node = root->rb_root.rb_node;
 968         struct work_atoms key = { .thread = thread };
 969
 970         while (node) {
 971                 struct work_atoms *atoms;
 972                 int cmp;
 973
 974                 atoms = container_of(node, struct work_atoms, node);
 975
 976                 cmp = thread_lat_cmp(sort_list, &key, atoms);
 977                 if (cmp > 0)
 978                         node = node->rb_left;
 979                 else if (cmp < 0)
 980                         node = node->rb_right;
 981                 else {
 982                         BUG_ON(thread != atoms->thread);
 983                         return atoms;
 984                 }
 985         }
 986         return NULL;
 987 }
 988
 989 static void
 990 __thread_latency_insert(struct rb_root_cached *root, struct work_atoms *data,
 991                          struct list_head *sort_list)
 992 {
 993         struct rb_node **new = &(root->rb_root.rb_node), *parent = NULL;
 994         bool leftmost = true;
 995
 996         while (*new) {
 997                 struct work_atoms *this;
 998                 int cmp;
 999
1000                 this = container_of(*new, struct work_atoms, node);
1001                 parent = *new;
1002
1003                 cmp = thread_lat_cmp(sort_list, data, this);
1004
1005                 if (cmp > 0)
1006                         new = &((*new)->rb_left);
1007                 else {
1008                         new = &((*new)->rb_right);
1009                         leftmost = false;
1010                 }
1011         }
1012
1013         rb_link_node(&data->node, parent, new);
1014         rb_insert_color_cached(&data->node, root, leftmost);
1015 }
1016
1017 static int thread_atoms_insert(struct perf_sched *sched, struct thread *thread)
1018 {
1019         struct work_atoms *atoms = zalloc(sizeof(*atoms));
1020         if (!atoms) {
1021                 pr_err("No memory at %s\n", __func__);
1022                 return -1;
1023         }
1024
1025         atoms->thread = thread__get(thread);
1026         INIT_LIST_HEAD(&atoms->work_list);
1027         __thread_latency_insert(&sched->atom_root, atoms, &sched->cmp_pid);
1028         return 0;
1029 }
1030
1031 static char sched_out_state(u64 prev_state)
1032 {
1033         const char *str = TASK_STATE_TO_CHAR_STR;
1034
1035         return str[prev_state];
1036 }
1037
1038 static int
1039 add_sched_out_event(struct work_atoms *atoms,
1040                     char run_state,
1041                     u64 timestamp)
1042 {
1043         struct work_atom *atom = zalloc(sizeof(*atom));
1044         if (!atom) {
1045                 pr_err("Non memory at %s", __func__);
1046                 return -1;
1047         }
1048
1049         atom->sched_out_time = timestamp;
1050
1051         if (run_state == 'R') {
1052                 atom->state = THREAD_WAIT_CPU;
1053                 atom->wake_up_time = atom->sched_out_time;
1054         }
1055
1056         list_add_tail(&atom->list, &atoms->work_list);
1057         return 0;
1058 }
1059
1060 static void
1061 add_runtime_event(struct work_atoms *atoms, u64 delta,
1062                   u64 timestamp __maybe_unused)
1063 {
1064         struct work_atom *atom;
1065
1066         BUG_ON(list_empty(&atoms->work_list));
1067
1068         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1069
1070         atom->runtime += delta;
1071         atoms->total_runtime += delta;
1072 }
1073
1074 static void
1075 add_sched_in_event(struct work_atoms *atoms, u64 timestamp)
1076 {
1077         struct work_atom *atom;
1078         u64 delta;
1079
1080         if (list_empty(&atoms->work_list))
1081                 return;
1082
1083         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1084
1085         if (atom->state != THREAD_WAIT_CPU)
1086                 return;
1087
1088         if (timestamp < atom->wake_up_time) {
1089                 atom->state = THREAD_IGNORE;
1090                 return;
1091         }
1092
1093         atom->state = THREAD_SCHED_IN;
1094         atom->sched_in_time = timestamp;
1095
1096         delta = atom->sched_in_time - atom->wake_up_time;
1097         atoms->total_lat += delta;
1098         if (delta > atoms->max_lat) {
1099                 atoms->max_lat = delta;
1100                 atoms->max_lat_start = atom->wake_up_time;
1101                 atoms->max_lat_end = timestamp;
1102         }
1103         atoms->nb_atoms++;
1104 }
1105
1106 static int latency_switch_event(struct perf_sched *sched,
1107                                 struct evsel *evsel,
1108                                 struct perf_sample *sample,
1109                                 struct machine *machine)
1110 {
1111         const u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
1112                   next_pid = evsel__intval(evsel, sample, "next_pid");
1113         const u64 prev_state = evsel__intval(evsel, sample, "prev_state");
1114         struct work_atoms *out_events, *in_events;
1115         struct thread *sched_out, *sched_in;
1116         u64 timestamp0, timestamp = sample->time;
1117         int cpu = sample->cpu, err = -1;
1118         s64 delta;
1119
1120         BUG_ON(cpu >= MAX_CPUS || cpu < 0);
1121
1122         timestamp0 = sched->cpu_last_switched[cpu];
1123         sched->cpu_last_switched[cpu] = timestamp;
1124         if (timestamp0)
1125                 delta = timestamp - timestamp0;
1126         else
1127                 delta = 0;
1128
1129         if (delta < 0) {
1130                 pr_err("hm, delta: %" PRIu64 " < 0 ?\n", delta);
1131                 return -1;
1132         }
1133
1134         sched_out = machine__findnew_thread(machine, -1, prev_pid);
1135         sched_in = machine__findnew_thread(machine, -1, next_pid);
1136         if (sched_out == NULL || sched_in == NULL)
1137                 goto out_put;
1138
1139         out_events = thread_atoms_search(&sched->atom_root, sched_out, &sched->cmp_pid);
1140         if (!out_events) {
1141                 if (thread_atoms_insert(sched, sched_out))
1142                         goto out_put;
1143                 out_events = thread_atoms_search(&sched->atom_root, sched_out, &sched->cmp_pid);
1144                 if (!out_events) {
1145                         pr_err("out-event: Internal tree error");
1146                         goto out_put;
1147                 }
1148         }
1149         if (add_sched_out_event(out_events, sched_out_state(prev_state), timestamp))
1150                 return -1;
1151
1152         in_events = thread_atoms_search(&sched->atom_root, sched_in, &sched->cmp_pid);
1153         if (!in_events) {
1154                 if (thread_atoms_insert(sched, sched_in))
1155                         goto out_put;
1156                 in_events = thread_atoms_search(&sched->atom_root, sched_in, &sched->cmp_pid);
1157                 if (!in_events) {
1158                         pr_err("in-event: Internal tree error");
1159                         goto out_put;
1160                 }
1161                 /*
1162                  * Take came in we have not heard about yet,
1163                  * add in an initial atom in runnable state:
1164                  */
1165                 if (add_sched_out_event(in_events, 'R', timestamp))
1166                         goto out_put;
1167         }
1168         add_sched_in_event(in_events, timestamp);
1169         err = 0;
1170 out_put:
1171         thread__put(sched_out);
1172         thread__put(sched_in);
1173         return err;
1174 }
1175
1176 static int latency_runtime_event(struct perf_sched *sched,
1177                                  struct evsel *evsel,
1178                                  struct perf_sample *sample,
1179                                  struct machine *machine)
1180 {
1181         const u32 pid      = evsel__intval(evsel, sample, "pid");
1182         const u64 runtime  = evsel__intval(evsel, sample, "runtime");
1183         struct thread *thread = machine__findnew_thread(machine, -1, pid);
1184         struct work_atoms *atoms = thread_atoms_search(&sched->atom_root, thread, &sched->cmp_pid);
1185         u64 timestamp = sample->time;
1186         int cpu = sample->cpu, err = -1;
1187
1188         if (thread == NULL)
1189                 return -1;
1190
1191         BUG_ON(cpu >= MAX_CPUS || cpu < 0);
1192         if (!atoms) {
1193                 if (thread_atoms_insert(sched, thread))
1194                         goto out_put;
1195                 atoms = thread_atoms_search(&sched->atom_root, thread, &sched->cmp_pid);
1196                 if (!atoms) {
1197                         pr_err("in-event: Internal tree error");
1198                         goto out_put;
1199                 }
1200                 if (add_sched_out_event(atoms, 'R', timestamp))
1201                         goto out_put;
1202         }
1203
1204         add_runtime_event(atoms, runtime, timestamp);
1205         err = 0;
1206 out_put:
1207         thread__put(thread);
1208         return err;
1209 }
1210
1211 static int latency_wakeup_event(struct perf_sched *sched,
1212                                 struct evsel *evsel,
1213                                 struct perf_sample *sample,
1214                                 struct machine *machine)
1215 {
1216         const u32 pid     = evsel__intval(evsel, sample, "pid");
1217         struct work_atoms *atoms;
1218         struct work_atom *atom;
1219         struct thread *wakee;
1220         u64 timestamp = sample->time;
1221         int err = -1;
1222
1223         wakee = machine__findnew_thread(machine, -1, pid);
1224         if (wakee == NULL)
1225                 return -1;
1226         atoms = thread_atoms_search(&sched->atom_root, wakee, &sched->cmp_pid);
1227         if (!atoms) {
1228                 if (thread_atoms_insert(sched, wakee))
1229                         goto out_put;
1230                 atoms = thread_atoms_search(&sched->atom_root, wakee, &sched->cmp_pid);
1231                 if (!atoms) {
1232                         pr_err("wakeup-event: Internal tree error");
1233                         goto out_put;
1234                 }
1235                 if (add_sched_out_event(atoms, 'S', timestamp))
1236                         goto out_put;
1237         }
1238
1239         BUG_ON(list_empty(&atoms->work_list));
1240
1241         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1242
1243         /*
1244          * As we do not guarantee the wakeup event happens when
1245          * task is out of run queue, also may happen when task is
1246          * on run queue and wakeup only change ->state to TASK_RUNNING,
1247          * then we should not set the ->wake_up_time when wake up a
1248          * task which is on run queue.
1249          *
1250          * You WILL be missing events if you've recorded only
1251          * one CPU, or are only looking at only one, so don't
1252          * skip in this case.
1253          */
1254         if (sched->profile_cpu == -1 && atom->state != THREAD_SLEEPING)
1255                 goto out_ok;
1256
1257         sched->nr_timestamps++;
1258         if (atom->sched_out_time > timestamp) {
1259                 sched->nr_unordered_timestamps++;
1260                 goto out_ok;
1261         }
1262
1263         atom->state = THREAD_WAIT_CPU;
1264         atom->wake_up_time = timestamp;
1265 out_ok:
1266         err = 0;
1267 out_put:
1268         thread__put(wakee);
1269         return err;
1270 }
1271
1272 static int latency_migrate_task_event(struct perf_sched *sched,
1273                                       struct evsel *evsel,
1274                                       struct perf_sample *sample,
1275                                       struct machine *machine)
1276 {
1277         const u32 pid = evsel__intval(evsel, sample, "pid");
1278         u64 timestamp = sample->time;
1279         struct work_atoms *atoms;
1280         struct work_atom *atom;
1281         struct thread *migrant;
1282         int err = -1;
1283
1284         /*
1285          * Only need to worry about migration when profiling one CPU.
1286          */
1287         if (sched->profile_cpu == -1)
1288                 return 0;
1289
1290         migrant = machine__findnew_thread(machine, -1, pid);
1291         if (migrant == NULL)
1292                 return -1;
1293         atoms = thread_atoms_search(&sched->atom_root, migrant, &sched->cmp_pid);
1294         if (!atoms) {
1295                 if (thread_atoms_insert(sched, migrant))
1296                         goto out_put;
1297                 register_pid(sched, migrant->tid, thread__comm_str(migrant));
1298                 atoms = thread_atoms_search(&sched->atom_root, migrant, &sched->cmp_pid);
1299                 if (!atoms) {
1300                         pr_err("migration-event: Internal tree error");
1301                         goto out_put;
1302                 }
1303                 if (add_sched_out_event(atoms, 'R', timestamp))
1304                         goto out_put;
1305         }
1306
1307         BUG_ON(list_empty(&atoms->work_list));
1308
1309         atom = list_entry(atoms->work_list.prev, struct work_atom, list);
1310         atom->sched_in_time = atom->sched_out_time = atom->wake_up_time = timestamp;
1311
1312         sched->nr_timestamps++;
1313
1314         if (atom->sched_out_time > timestamp)
1315                 sched->nr_unordered_timestamps++;
1316         err = 0;
1317 out_put:
1318         thread__put(migrant);
1319         return err;
1320 }
1321
1322 static void output_lat_thread(struct perf_sched *sched, struct work_atoms *work_list)
1323 {
1324         int i;
1325         int ret;
1326         u64 avg;
1327         char max_lat_start[32], max_lat_end[32];
1328
1329         if (!work_list->nb_atoms)
1330                 return;
1331         /*
1332          * Ignore idle threads:
1333          */
1334         if (!strcmp(thread__comm_str(work_list->thread), "swapper"))
1335                 return;
1336
1337         sched->all_runtime += work_list->total_runtime;
1338         sched->all_count   += work_list->nb_atoms;
1339
1340         if (work_list->num_merged > 1)
1341                 ret = printf("  %s:(%d) ", thread__comm_str(work_list->thread), work_list->num_merged);
1342         else
1343                 ret = printf("  %s:%d ", thread__comm_str(work_list->thread), work_list->thread->tid);
1344
1345         for (i = 0; i < 24 - ret; i++)
1346                 printf(" ");
1347
1348         avg = work_list->total_lat / work_list->nb_atoms;
1349         timestamp__scnprintf_usec(work_list->max_lat_start, max_lat_start, sizeof(max_lat_start));
1350         timestamp__scnprintf_usec(work_list->max_lat_end, max_lat_end, sizeof(max_lat_end));
1351
1352         printf("|%11.3f ms |%9" PRIu64 " | avg:%8.3f ms | max:%8.3f ms | max start: %12s s | max end: %12s s\n",
1353               (double)work_list->total_runtime / NSEC_PER_MSEC,
1354                  work_list->nb_atoms, (double)avg / NSEC_PER_MSEC,
1355                  (double)work_list->max_lat / NSEC_PER_MSEC,
1356                  max_lat_start, max_lat_end);
1357 }
1358
1359 static int pid_cmp(struct work_atoms *l, struct work_atoms *r)
1360 {
1361         if (l->thread == r->thread)
1362                 return 0;
1363         if (l->thread->tid < r->thread->tid)
1364                 return -1;
1365         if (l->thread->tid > r->thread->tid)
1366                 return 1;
1367         return (int)(l->thread - r->thread);
1368 }
1369
1370 static int avg_cmp(struct work_atoms *l, struct work_atoms *r)
1371 {
1372         u64 avgl, avgr;
1373
1374         if (!l->nb_atoms)
1375                 return -1;
1376
1377         if (!r->nb_atoms)
1378                 return 1;
1379
1380         avgl = l->total_lat / l->nb_atoms;
1381         avgr = r->total_lat / r->nb_atoms;
1382
1383         if (avgl < avgr)
1384                 return -1;
1385         if (avgl > avgr)
1386                 return 1;
1387
1388         return 0;
1389 }
1390
1391 static int max_cmp(struct work_atoms *l, struct work_atoms *r)
1392 {
1393         if (l->max_lat < r->max_lat)
1394                 return -1;
1395         if (l->max_lat > r->max_lat)
1396                 return 1;
1397
1398         return 0;
1399 }
1400
1401 static int switch_cmp(struct work_atoms *l, struct work_atoms *r)
1402 {
1403         if (l->nb_atoms < r->nb_atoms)
1404                 return -1;
1405         if (l->nb_atoms > r->nb_atoms)
1406                 return 1;
1407
1408         return 0;
1409 }
1410
1411 static int runtime_cmp(struct work_atoms *l, struct work_atoms *r)
1412 {
1413         if (l->total_runtime < r->total_runtime)
1414                 return -1;
1415         if (l->total_runtime > r->total_runtime)
1416                 return 1;
1417
1418         return 0;
1419 }
1420
1421 static int sort_dimension__add(const char *tok, struct list_head *list)
1422 {
1423         size_t i;
1424         static struct sort_dimension avg_sort_dimension = {
1425                 .name = "avg",
1426                 .cmp  = avg_cmp,
1427         };
1428         static struct sort_dimension max_sort_dimension = {
1429                 .name = "max",
1430                 .cmp  = max_cmp,
1431         };
1432         static struct sort_dimension pid_sort_dimension = {
1433                 .name = "pid",
1434                 .cmp  = pid_cmp,
1435         };
1436         static struct sort_dimension runtime_sort_dimension = {
1437                 .name = "runtime",
1438                 .cmp  = runtime_cmp,
1439         };
1440         static struct sort_dimension switch_sort_dimension = {
1441                 .name = "switch",
1442                 .cmp  = switch_cmp,
1443         };
1444         struct sort_dimension *available_sorts[] = {
1445                 &pid_sort_dimension,
1446                 &avg_sort_dimension,
1447                 &max_sort_dimension,
1448                 &switch_sort_dimension,
1449                 &runtime_sort_dimension,
1450         };
1451
1452         for (i = 0; i < ARRAY_SIZE(available_sorts); i++) {
1453                 if (!strcmp(available_sorts[i]->name, tok)) {
1454                         list_add_tail(&available_sorts[i]->list, list);
1455
1456                         return 0;
1457                 }
1458         }
1459
1460         return -1;
1461 }
1462
1463 static void perf_sched__sort_lat(struct perf_sched *sched)
1464 {
1465         struct rb_node *node;
1466         struct rb_root_cached *root = &sched->atom_root;
1467 again:
1468         for (;;) {
1469                 struct work_atoms *data;
1470                 node = rb_first_cached(root);
1471                 if (!node)
1472                         break;
1473
1474                 rb_erase_cached(node, root);
1475                 data = rb_entry(node, struct work_atoms, node);
1476                 __thread_latency_insert(&sched->sorted_atom_root, data, &sched->sort_list);
1477         }
1478         if (root == &sched->atom_root) {
1479                 root = &sched->merged_atom_root;
1480                 goto again;
1481         }
1482 }
1483
1484 static int process_sched_wakeup_event(struct perf_tool *tool,
1485                                       struct evsel *evsel,
1486                                       struct perf_sample *sample,
1487                                       struct machine *machine)
1488 {
1489         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
1490
1491         if (sched->tp_handler->wakeup_event)
1492                 return sched->tp_handler->wakeup_event(sched, evsel, sample, machine);
1493
1494         return 0;
1495 }
1496
1497 union map_priv {
1498         void    *ptr;
1499         bool     color;
1500 };
1501
1502 static bool thread__has_color(struct thread *thread)
1503 {
1504         union map_priv priv = {
1505                 .ptr = thread__priv(thread),
1506         };
1507
1508         return priv.color;
1509 }
1510
1511 static struct thread*
1512 map__findnew_thread(struct perf_sched *sched, struct machine *machine, pid_t pid, pid_t tid)
1513 {
1514         struct thread *thread = machine__findnew_thread(machine, pid, tid);
1515         union map_priv priv = {
1516                 .color = false,
1517         };
1518
1519         if (!sched->map.color_pids || !thread || thread__priv(thread))
1520                 return thread;
1521
1522         if (thread_map__has(sched->map.color_pids, tid))
1523                 priv.color = true;
1524
1525         thread__set_priv(thread, priv.ptr);
1526         return thread;
1527 }
1528
1529 static int map_switch_event(struct perf_sched *sched, struct evsel *evsel,
1530                             struct perf_sample *sample, struct machine *machine)
1531 {
1532         const u32 next_pid = evsel__intval(evsel, sample, "next_pid");
1533         struct thread *sched_in;
1534         struct thread_runtime *tr;
1535         int new_shortname;
1536         u64 timestamp0, timestamp = sample->time;
1537         s64 delta;
1538         int i;
1539         struct perf_cpu this_cpu = {
1540                 .cpu = sample->cpu,
1541         };
1542         int cpus_nr;
1543         bool new_cpu = false;
1544         const char *color = PERF_COLOR_NORMAL;
1545         char stimestamp[32];
1546
1547         BUG_ON(this_cpu.cpu >= MAX_CPUS || this_cpu.cpu < 0);
1548
1549         if (this_cpu.cpu > sched->max_cpu.cpu)
1550                 sched->max_cpu = this_cpu;
1551
1552         if (sched->map.comp) {
1553                 cpus_nr = bitmap_weight(sched->map.comp_cpus_mask, MAX_CPUS);
1554                 if (!test_and_set_bit(this_cpu.cpu, sched->map.comp_cpus_mask)) {
1555                         sched->map.comp_cpus[cpus_nr++] = this_cpu;
1556                         new_cpu = true;
1557                 }
1558         } else
1559                 cpus_nr = sched->max_cpu.cpu;
1560
1561         timestamp0 = sched->cpu_last_switched[this_cpu.cpu];
1562         sched->cpu_last_switched[this_cpu.cpu] = timestamp;
1563         if (timestamp0)
1564                 delta = timestamp - timestamp0;
1565         else
1566                 delta = 0;
1567
1568         if (delta < 0) {
1569                 pr_err("hm, delta: %" PRIu64 " < 0 ?\n", delta);
1570                 return -1;
1571         }
1572
1573         sched_in = map__findnew_thread(sched, machine, -1, next_pid);
1574         if (sched_in == NULL)
1575                 return -1;
1576
1577         tr = thread__get_runtime(sched_in);
1578         if (tr == NULL) {
1579                 thread__put(sched_in);
1580                 return -1;
1581         }
1582
1583         sched->curr_thread[this_cpu.cpu] = thread__get(sched_in);
1584
1585         printf("  ");
1586
1587         new_shortname = 0;
1588         if (!tr->shortname[0]) {
1589                 if (!strcmp(thread__comm_str(sched_in), "swapper")) {
1590                         /*
1591                          * Don't allocate a letter-number for swapper:0
1592                          * as a shortname. Instead, we use '.' for it.
1593                          */
1594                         tr->shortname[0] = '.';
1595                         tr->shortname[1] = ' ';
1596                 } else {
1597                         tr->shortname[0] = sched->next_shortname1;
1598                         tr->shortname[1] = sched->next_shortname2;
1599
1600                         if (sched->next_shortname1 < 'Z') {
1601                                 sched->next_shortname1++;
1602                         } else {
1603                                 sched->next_shortname1 = 'A';
1604                                 if (sched->next_shortname2 < '9')
1605                                         sched->next_shortname2++;
1606                                 else
1607                                         sched->next_shortname2 = '0';
1608                         }
1609                 }
1610                 new_shortname = 1;
1611         }
1612
1613         for (i = 0; i < cpus_nr; i++) {
1614                 struct perf_cpu cpu = {
1615                         .cpu = sched->map.comp ? sched->map.comp_cpus[i].cpu : i,
1616                 };
1617                 struct thread *curr_thread = sched->curr_thread[cpu.cpu];
1618                 struct thread_runtime *curr_tr;
1619                 const char *pid_color = color;
1620                 const char *cpu_color = color;
1621
1622                 if (curr_thread && thread__has_color(curr_thread))
1623                         pid_color = COLOR_PIDS;
1624
1625                 if (sched->map.cpus && !perf_cpu_map__has(sched->map.cpus, cpu))
1626                         continue;
1627
1628                 if (sched->map.color_cpus && perf_cpu_map__has(sched->map.color_cpus, cpu))
1629                         cpu_color = COLOR_CPUS;
1630
1631                 if (cpu.cpu != this_cpu.cpu)
1632                         color_fprintf(stdout, color, " ");
1633                 else
1634                         color_fprintf(stdout, cpu_color, "*");
1635
1636                 if (sched->curr_thread[cpu.cpu]) {
1637                         curr_tr = thread__get_runtime(sched->curr_thread[cpu.cpu]);
1638                         if (curr_tr == NULL) {
1639                                 thread__put(sched_in);
1640                                 return -1;
1641                         }
1642                         color_fprintf(stdout, pid_color, "%2s ", curr_tr->shortname);
1643                 } else
1644                         color_fprintf(stdout, color, "   ");
1645         }
1646
1647         if (sched->map.cpus && !perf_cpu_map__has(sched->map.cpus, this_cpu))
1648                 goto out;
1649
1650         timestamp__scnprintf_usec(timestamp, stimestamp, sizeof(stimestamp));
1651         color_fprintf(stdout, color, "  %12s secs ", stimestamp);
1652         if (new_shortname || tr->comm_changed || (verbose > 0 && sched_in->tid)) {
1653                 const char *pid_color = color;
1654
1655                 if (thread__has_color(sched_in))
1656                         pid_color = COLOR_PIDS;
1657
1658                 color_fprintf(stdout, pid_color, "%s => %s:%d",
1659                        tr->shortname, thread__comm_str(sched_in), sched_in->tid);
1660                 tr->comm_changed = false;
1661         }
1662
1663         if (sched->map.comp && new_cpu)
1664                 color_fprintf(stdout, color, " (CPU %d)", this_cpu);
1665
1666 out:
1667         color_fprintf(stdout, color, "\n");
1668
1669         thread__put(sched_in);
1670
1671         return 0;
1672 }
1673
1674 static int process_sched_switch_event(struct perf_tool *tool,
1675                                       struct evsel *evsel,
1676                                       struct perf_sample *sample,
1677                                       struct machine *machine)
1678 {
1679         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
1680         int this_cpu = sample->cpu, err = 0;
1681         u32 prev_pid = evsel__intval(evsel, sample, "prev_pid"),
1682             next_pid = evsel__intval(evsel, sample, "next_pid");
1683
1684         if (sched->curr_pid[this_cpu] != (u32)-1) {
1685                 /*
1686                  * Are we trying to switch away a PID that is
1687                  * not current?
1688                  */
1689                 if (sched->curr_pid[this_cpu] != prev_pid)
1690                         sched->nr_context_switch_bugs++;
1691         }
1692
1693         if (sched->tp_handler->switch_event)
1694                 err = sched->tp_handler->switch_event(sched, evsel, sample, machine);
1695
1696         sched->curr_pid[this_cpu] = next_pid;
1697         return err;
1698 }
1699
1700 static int process_sched_runtime_event(struct perf_tool *tool,
1701                                        struct evsel *evsel,
1702                                        struct perf_sample *sample,
1703                                        struct machine *machine)
1704 {
1705         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
1706
1707         if (sched->tp_handler->runtime_event)
1708                 return sched->tp_handler->runtime_event(sched, evsel, sample, machine);
1709
1710         return 0;
1711 }
1712
1713 static int perf_sched__process_fork_event(struct perf_tool *tool,
1714                                           union perf_event *event,
1715                                           struct perf_sample *sample,
1716                                           struct machine *machine)
1717 {
1718         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
1719
1720         /* run the fork event through the perf machinery */
1721         perf_event__process_fork(tool, event, sample, machine);
1722
1723         /* and then run additional processing needed for this command */
1724         if (sched->tp_handler->fork_event)
1725                 return sched->tp_handler->fork_event(sched, event, machine);
1726
1727         return 0;
1728 }
1729
1730 static int process_sched_migrate_task_event(struct perf_tool *tool,
1731                                             struct evsel *evsel,
1732                                             struct perf_sample *sample,
1733                                             struct machine *machine)
1734 {
1735         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
1736
1737         if (sched->tp_handler->migrate_task_event)
1738                 return sched->tp_handler->migrate_task_event(sched, evsel, sample, machine);
1739
1740         return 0;
1741 }
1742
1743 typedef int (*tracepoint_handler)(struct perf_tool *tool,
1744                                   struct evsel *evsel,
1745                                   struct perf_sample *sample,
1746                                   struct machine *machine);
1747
1748 static int perf_sched__process_tracepoint_sample(struct perf_tool *tool __maybe_unused,
1749                                                  union perf_event *event __maybe_unused,
1750                                                  struct perf_sample *sample,
1751                                                  struct evsel *evsel,
1752                                                  struct machine *machine)
1753 {
1754         int err = 0;
1755
1756         if (evsel->handler != NULL) {
1757                 tracepoint_handler f = evsel->handler;
1758                 err = f(tool, evsel, sample, machine);
1759         }
1760
1761         return err;
1762 }
1763
1764 static int perf_sched__process_comm(struct perf_tool *tool __maybe_unused,
1765                                     union perf_event *event,
1766                                     struct perf_sample *sample,
1767                                     struct machine *machine)
1768 {
1769         struct thread *thread;
1770         struct thread_runtime *tr;
1771         int err;
1772
1773         err = perf_event__process_comm(tool, event, sample, machine);
1774         if (err)
1775                 return err;
1776
1777         thread = machine__find_thread(machine, sample->pid, sample->tid);
1778         if (!thread) {
1779                 pr_err("Internal error: can't find thread\n");
1780                 return -1;
1781         }
1782
1783         tr = thread__get_runtime(thread);
1784         if (tr == NULL) {
1785                 thread__put(thread);
1786                 return -1;
1787         }
1788
1789         tr->comm_changed = true;
1790         thread__put(thread);
1791
1792         return 0;
1793 }
1794
1795 static int perf_sched__read_events(struct perf_sched *sched)
1796 {
1797         const struct evsel_str_handler handlers[] = {
1798                 { "sched:sched_switch",       process_sched_switch_event, },
1799                 { "sched:sched_stat_runtime", process_sched_runtime_event, },
1800                 { "sched:sched_wakeup",       process_sched_wakeup_event, },
1801                 { "sched:sched_wakeup_new",   process_sched_wakeup_event, },
1802                 { "sched:sched_migrate_task", process_sched_migrate_task_event, },
1803         };
1804         struct perf_session *session;
1805         struct perf_data data = {
1806                 .path  = input_name,
1807                 .mode  = PERF_DATA_MODE_READ,
1808                 .force = sched->force,
1809         };
1810         int rc = -1;
1811
1812         session = perf_session__new(&data, &sched->tool);
1813         if (IS_ERR(session)) {
1814                 pr_debug("Error creating perf session");
1815                 return PTR_ERR(session);
1816         }
1817
1818         symbol__init(&session->header.env);
1819
1820         if (perf_session__set_tracepoints_handlers(session, handlers))
1821                 goto out_delete;
1822
1823         if (perf_session__has_traces(session, "record -R")) {
1824                 int err = perf_session__process_events(session);
1825                 if (err) {
1826                         pr_err("Failed to process events, error %d", err);
1827                         goto out_delete;
1828                 }
1829
1830                 sched->nr_events      = session->evlist->stats.nr_events[0];
1831                 sched->nr_lost_events = session->evlist->stats.total_lost;
1832                 sched->nr_lost_chunks = session->evlist->stats.nr_events[PERF_RECORD_LOST];
1833         }
1834
1835         rc = 0;
1836 out_delete:
1837         perf_session__delete(session);
1838         return rc;
1839 }
1840
1841 /*
1842  * scheduling times are printed as msec.usec
1843  */
1844 static inline void print_sched_time(unsigned long long nsecs, int width)
1845 {
1846         unsigned long msecs;
1847         unsigned long usecs;
1848
1849         msecs  = nsecs / NSEC_PER_MSEC;
1850         nsecs -= msecs * NSEC_PER_MSEC;
1851         usecs  = nsecs / NSEC_PER_USEC;
1852         printf("%*lu.%03lu ", width, msecs, usecs);
1853 }
1854
1855 /*
1856  * returns runtime data for event, allocating memory for it the
1857  * first time it is used.
1858  */
1859 static struct evsel_runtime *evsel__get_runtime(struct evsel *evsel)
1860 {
1861         struct evsel_runtime *r = evsel->priv;
1862
1863         if (r == NULL) {
1864                 r = zalloc(sizeof(struct evsel_runtime));
1865                 evsel->priv = r;
1866         }
1867
1868         return r;
1869 }
1870
1871 /*
1872  * save last time event was seen per cpu
1873  */
1874 static void evsel__save_time(struct evsel *evsel, u64 timestamp, u32 cpu)
1875 {
1876         struct evsel_runtime *r = evsel__get_runtime(evsel);
1877
1878         if (r == NULL)
1879                 return;
1880
1881         if ((cpu >= r->ncpu) || (r->last_time == NULL)) {
1882                 int i, n = __roundup_pow_of_two(cpu+1);
1883                 void *p = r->last_time;
1884
1885                 p = realloc(r->last_time, n * sizeof(u64));
1886                 if (!p)
1887                         return;
1888
1889                 r->last_time = p;
1890                 for (i = r->ncpu; i < n; ++i)
1891                         r->last_time[i] = (u64) 0;
1892
1893                 r->ncpu = n;
1894         }
1895
1896         r->last_time[cpu] = timestamp;
1897 }
1898
1899 /* returns last time this event was seen on the given cpu */
1900 static u64 evsel__get_time(struct evsel *evsel, u32 cpu)
1901 {
1902         struct evsel_runtime *r = evsel__get_runtime(evsel);
1903
1904         if ((r == NULL) || (r->last_time == NULL) || (cpu >= r->ncpu))
1905                 return 0;
1906
1907         return r->last_time[cpu];
1908 }
1909
1910 static int comm_width = 30;
1911
1912 static char *timehist_get_commstr(struct thread *thread)
1913 {
1914         static char str[32];
1915         const char *comm = thread__comm_str(thread);
1916         pid_t tid = thread->tid;
1917         pid_t pid = thread->pid_;
1918         int n;
1919
1920         if (pid == 0)
1921                 n = scnprintf(str, sizeof(str), "%s", comm);
1922
1923         else if (tid != pid)
1924                 n = scnprintf(str, sizeof(str), "%s[%d/%d]", comm, tid, pid);
1925
1926         else
1927                 n = scnprintf(str, sizeof(str), "%s[%d]", comm, tid);
1928
1929         if (n > comm_width)
1930                 comm_width = n;
1931
1932         return str;
1933 }
1934
1935 static void timehist_header(struct perf_sched *sched)
1936 {
1937         u32 ncpus = sched->max_cpu.cpu + 1;
1938         u32 i, j;
1939
1940         printf("%15s %6s ", "time", "cpu");
1941
1942         if (sched->show_cpu_visual) {
1943                 printf(" ");
1944                 for (i = 0, j = 0; i < ncpus; ++i) {
1945                         printf("%x", j++);
1946                         if (j > 15)
1947                                 j = 0;
1948                 }
1949                 printf(" ");
1950         }
1951
1952         printf(" %-*s  %9s  %9s  %9s", comm_width,
1953                 "task name", "wait time", "sch delay", "run time");
1954
1955         if (sched->show_state)
1956                 printf("  %s", "state");
1957
1958         printf("\n");
1959
1960         /*
1961          * units row
1962          */
1963         printf("%15s %-6s ", "", "");
1964
1965         if (sched->show_cpu_visual)
1966                 printf(" %*s ", ncpus, "");
1967
1968         printf(" %-*s  %9s  %9s  %9s", comm_width,
1969                "[tid/pid]", "(msec)", "(msec)", "(msec)");
1970
1971         if (sched->show_state)
1972                 printf("  %5s", "");
1973
1974         printf("\n");
1975
1976         /*
1977          * separator
1978          */
1979         printf("%.15s %.6s ", graph_dotted_line, graph_dotted_line);
1980
1981         if (sched->show_cpu_visual)
1982                 printf(" %.*s ", ncpus, graph_dotted_line);
1983
1984         printf(" %.*s  %.9s  %.9s  %.9s", comm_width,
1985                 graph_dotted_line, graph_dotted_line, graph_dotted_line,
1986                 graph_dotted_line);
1987
1988         if (sched->show_state)
1989                 printf("  %.5s", graph_dotted_line);
1990
1991         printf("\n");
1992 }
1993
1994 static char task_state_char(struct thread *thread, int state)
1995 {
1996         static const char state_to_char[] = TASK_STATE_TO_CHAR_STR;
1997         unsigned bit = state ? ffs(state) : 0;
1998
1999         /* 'I' for idle */
2000         if (thread->tid == 0)
2001                 return 'I';
2002
2003         return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?';
2004 }
2005
2006 static void timehist_print_sample(struct perf_sched *sched,
2007                                   struct evsel *evsel,
2008                                   struct perf_sample *sample,
2009                                   struct addr_location *al,
2010                                   struct thread *thread,
2011                                   u64 t, int state)
2012 {
2013         struct thread_runtime *tr = thread__priv(thread);
2014         const char *next_comm = evsel__strval(evsel, sample, "next_comm");
2015         const u32 next_pid = evsel__intval(evsel, sample, "next_pid");
2016         u32 max_cpus = sched->max_cpu.cpu + 1;
2017         char tstr[64];
2018         char nstr[30];
2019         u64 wait_time;
2020
2021         if (cpu_list && !test_bit(sample->cpu, cpu_bitmap))
2022                 return;
2023
2024         timestamp__scnprintf_usec(t, tstr, sizeof(tstr));
2025         printf("%15s [%04d] ", tstr, sample->cpu);
2026
2027         if (sched->show_cpu_visual) {
2028                 u32 i;
2029                 char c;
2030
2031                 printf(" ");
2032                 for (i = 0; i < max_cpus; ++i) {
2033                         /* flag idle times with 'i'; others are sched events */
2034                         if (i == sample->cpu)
2035                                 c = (thread->tid == 0) ? 'i' : 's';
2036                         else
2037                                 c = ' ';
2038                         printf("%c", c);
2039                 }
2040                 printf(" ");
2041         }
2042
2043         printf(" %-*s ", comm_width, timehist_get_commstr(thread));
2044
2045         wait_time = tr->dt_sleep + tr->dt_iowait + tr->dt_preempt;
2046         print_sched_time(wait_time, 6);
2047
2048         print_sched_time(tr->dt_delay, 6);
2049         print_sched_time(tr->dt_run, 6);
2050
2051         if (sched->show_state)
2052                 printf(" %5c ", task_state_char(thread, state));
2053
2054         if (sched->show_next) {
2055                 snprintf(nstr, sizeof(nstr), "next: %s[%d]", next_comm, next_pid);
2056                 printf(" %-*s", comm_width, nstr);
2057         }
2058
2059         if (sched->show_wakeups && !sched->show_next)
2060                 printf("  %-*s", comm_width, "");
2061
2062         if (thread->tid == 0)
2063                 goto out;
2064
2065         if (sched->show_callchain)
2066                 printf("  ");
2067
2068         sample__fprintf_sym(sample, al, 0,
2069                             EVSEL__PRINT_SYM | EVSEL__PRINT_ONELINE |
2070                             EVSEL__PRINT_CALLCHAIN_ARROW |
2071                             EVSEL__PRINT_SKIP_IGNORED,
2072                             &callchain_cursor, symbol_conf.bt_stop_list,  stdout);
2073
2074 out:
2075         printf("\n");
2076 }
2077
2078 /*
2079  * Explanation of delta-time stats:
2080  *
2081  *            t = time of current schedule out event
2082  *        tprev = time of previous sched out event
2083  *                also time of schedule-in event for current task
2084  *    last_time = time of last sched change event for current task
2085  *                (i.e, time process was last scheduled out)
2086  * ready_to_run = time of wakeup for current task
2087  *
2088  * -----|------------|------------|------------|------
2089  *    last         ready        tprev          t
2090  *    time         to run
2091  *
2092  *      |-------- dt_wait --------|
2093  *                   |- dt_delay -|-- dt_run --|
2094  *
2095  *   dt_run = run time of current task
2096  *  dt_wait = time between last schedule out event for task and tprev
2097  *            represents time spent off the cpu
2098  * dt_delay = time between wakeup and schedule-in of task
2099  */
2100
2101 static void timehist_update_runtime_stats(struct thread_runtime *r,
2102                                          u64 t, u64 tprev)
2103 {
2104         r->dt_delay   = 0;
2105         r->dt_sleep   = 0;
2106         r->dt_iowait  = 0;
2107         r->dt_preempt = 0;
2108         r->dt_run     = 0;
2109
2110         if (tprev) {
2111                 r->dt_run = t - tprev;
2112                 if (r->ready_to_run) {
2113                         if (r->ready_to_run > tprev)
2114                                 pr_debug("time travel: wakeup time for task > previous sched_switch event\n");
2115                         else
2116                                 r->dt_delay = tprev - r->ready_to_run;
2117                 }
2118
2119                 if (r->last_time > tprev)
2120                         pr_debug("time travel: last sched out time for task > previous sched_switch event\n");
2121                 else if (r->last_time) {
2122                         u64 dt_wait = tprev - r->last_time;
2123
2124                         if (r->last_state == TASK_RUNNING)
2125                                 r->dt_preempt = dt_wait;
2126                         else if (r->last_state == TASK_UNINTERRUPTIBLE)
2127                                 r->dt_iowait = dt_wait;
2128                         else
2129                                 r->dt_sleep = dt_wait;
2130                 }
2131         }
2132
2133         update_stats(&r->run_stats, r->dt_run);
2134
2135         r->total_run_time     += r->dt_run;
2136         r->total_delay_time   += r->dt_delay;
2137         r->total_sleep_time   += r->dt_sleep;
2138         r->total_iowait_time  += r->dt_iowait;
2139         r->total_preempt_time += r->dt_preempt;
2140 }
2141
2142 static bool is_idle_sample(struct perf_sample *sample,
2143                            struct evsel *evsel)
2144 {
2145         /* pid 0 == swapper == idle task */
2146         if (strcmp(evsel__name(evsel), "sched:sched_switch") == 0)
2147                 return evsel__intval(evsel, sample, "prev_pid") == 0;
2148
2149         return sample->pid == 0;
2150 }
2151
2152 static void save_task_callchain(struct perf_sched *sched,
2153                                 struct perf_sample *sample,
2154                                 struct evsel *evsel,
2155                                 struct machine *machine)
2156 {
2157         struct callchain_cursor *cursor = &callchain_cursor;
2158         struct thread *thread;
2159
2160         /* want main thread for process - has maps */
2161         thread = machine__findnew_thread(machine, sample->pid, sample->pid);
2162         if (thread == NULL) {
2163                 pr_debug("Failed to get thread for pid %d.\n", sample->pid);
2164                 return;
2165         }
2166
2167         if (!sched->show_callchain || sample->callchain == NULL)
2168                 return;
2169
2170         if (thread__resolve_callchain(thread, cursor, evsel, sample,
2171                                       NULL, NULL, sched->max_stack + 2) != 0) {
2172                 if (verbose > 0)
2173                         pr_err("Failed to resolve callchain. Skipping\n");
2174
2175                 return;
2176         }
2177
2178         callchain_cursor_commit(cursor);
2179
2180         while (true) {
2181                 struct callchain_cursor_node *node;
2182                 struct symbol *sym;
2183
2184                 node = callchain_cursor_current(cursor);
2185                 if (node == NULL)
2186                         break;
2187
2188                 sym = node->ms.sym;
2189                 if (sym) {
2190                         if (!strcmp(sym->name, "schedule") ||
2191                             !strcmp(sym->name, "__schedule") ||
2192                             !strcmp(sym->name, "preempt_schedule"))
2193                                 sym->ignore = 1;
2194                 }
2195
2196                 callchain_cursor_advance(cursor);
2197         }
2198 }
2199
2200 static int init_idle_thread(struct thread *thread)
2201 {
2202         struct idle_thread_runtime *itr;
2203
2204         thread__set_comm(thread, idle_comm, 0);
2205
2206         itr = zalloc(sizeof(*itr));
2207         if (itr == NULL)
2208                 return -ENOMEM;
2209
2210         init_stats(&itr->tr.run_stats);
2211         callchain_init(&itr->callchain);
2212         callchain_cursor_reset(&itr->cursor);
2213         thread__set_priv(thread, itr);
2214
2215         return 0;
2216 }
2217
2218 /*
2219  * Track idle stats per cpu by maintaining a local thread
2220  * struct for the idle task on each cpu.
2221  */
2222 static int init_idle_threads(int ncpu)
2223 {
2224         int i, ret;
2225
2226         idle_threads = zalloc(ncpu * sizeof(struct thread *));
2227         if (!idle_threads)
2228                 return -ENOMEM;
2229
2230         idle_max_cpu = ncpu;
2231
2232         /* allocate the actual thread struct if needed */
2233         for (i = 0; i < ncpu; ++i) {
2234                 idle_threads[i] = thread__new(0, 0);
2235                 if (idle_threads[i] == NULL)
2236                         return -ENOMEM;
2237
2238                 ret = init_idle_thread(idle_threads[i]);
2239                 if (ret < 0)
2240                         return ret;
2241         }
2242
2243         return 0;
2244 }
2245
2246 static void free_idle_threads(void)
2247 {
2248         int i;
2249
2250         if (idle_threads == NULL)
2251                 return;
2252
2253         for (i = 0; i < idle_max_cpu; ++i) {
2254                 if ((idle_threads[i]))
2255                         thread__delete(idle_threads[i]);
2256         }
2257
2258         free(idle_threads);
2259 }
2260
2261 static struct thread *get_idle_thread(int cpu)
2262 {
2263         /*
2264          * expand/allocate array of pointers to local thread
2265          * structs if needed
2266          */
2267         if ((cpu >= idle_max_cpu) || (idle_threads == NULL)) {
2268                 int i, j = __roundup_pow_of_two(cpu+1);
2269                 void *p;
2270
2271                 p = realloc(idle_threads, j * sizeof(struct thread *));
2272                 if (!p)
2273                         return NULL;
2274
2275                 idle_threads = (struct thread **) p;
2276                 for (i = idle_max_cpu; i < j; ++i)
2277                         idle_threads[i] = NULL;
2278
2279                 idle_max_cpu = j;
2280         }
2281
2282         /* allocate a new thread struct if needed */
2283         if (idle_threads[cpu] == NULL) {
2284                 idle_threads[cpu] = thread__new(0, 0);
2285
2286                 if (idle_threads[cpu]) {
2287                         if (init_idle_thread(idle_threads[cpu]) < 0)
2288                                 return NULL;
2289                 }
2290         }
2291
2292         return idle_threads[cpu];
2293 }
2294
2295 static void save_idle_callchain(struct perf_sched *sched,
2296                                 struct idle_thread_runtime *itr,
2297                                 struct perf_sample *sample)
2298 {
2299         if (!sched->show_callchain || sample->callchain == NULL)
2300                 return;
2301
2302         callchain_cursor__copy(&itr->cursor, &callchain_cursor);
2303 }
2304
2305 static struct thread *timehist_get_thread(struct perf_sched *sched,
2306                                           struct perf_sample *sample,
2307                                           struct machine *machine,
2308                                           struct evsel *evsel)
2309 {
2310         struct thread *thread;
2311
2312         if (is_idle_sample(sample, evsel)) {
2313                 thread = get_idle_thread(sample->cpu);
2314                 if (thread == NULL)
2315                         pr_err("Failed to get idle thread for cpu %d.\n", sample->cpu);
2316
2317         } else {
2318                 /* there were samples with tid 0 but non-zero pid */
2319                 thread = machine__findnew_thread(machine, sample->pid,
2320                                                  sample->tid ?: sample->pid);
2321                 if (thread == NULL) {
2322                         pr_debug("Failed to get thread for tid %d. skipping sample.\n",
2323                                  sample->tid);
2324                 }
2325
2326                 save_task_callchain(sched, sample, evsel, machine);
2327                 if (sched->idle_hist) {
2328                         struct thread *idle;
2329                         struct idle_thread_runtime *itr;
2330
2331                         idle = get_idle_thread(sample->cpu);
2332                         if (idle == NULL) {
2333                                 pr_err("Failed to get idle thread for cpu %d.\n", sample->cpu);
2334                                 return NULL;
2335                         }
2336
2337                         itr = thread__priv(idle);
2338                         if (itr == NULL)
2339                                 return NULL;
2340
2341                         itr->last_thread = thread;
2342
2343                         /* copy task callchain when entering to idle */
2344                         if (evsel__intval(evsel, sample, "next_pid") == 0)
2345                                 save_idle_callchain(sched, itr, sample);
2346                 }
2347         }
2348
2349         return thread;
2350 }
2351
2352 static bool timehist_skip_sample(struct perf_sched *sched,
2353                                  struct thread *thread,
2354                                  struct evsel *evsel,
2355                                  struct perf_sample *sample)
2356 {
2357         bool rc = false;
2358
2359         if (thread__is_filtered(thread)) {
2360                 rc = true;
2361                 sched->skipped_samples++;
2362         }
2363
2364         if (sched->idle_hist) {
2365                 if (strcmp(evsel__name(evsel), "sched:sched_switch"))
2366                         rc = true;
2367                 else if (evsel__intval(evsel, sample, "prev_pid") != 0 &&
2368                          evsel__intval(evsel, sample, "next_pid") != 0)
2369                         rc = true;
2370         }
2371
2372         return rc;
2373 }
2374
2375 static void timehist_print_wakeup_event(struct perf_sched *sched,
2376                                         struct evsel *evsel,
2377                                         struct perf_sample *sample,
2378                                         struct machine *machine,
2379                                         struct thread *awakened)
2380 {
2381         struct thread *thread;
2382         char tstr[64];
2383
2384         thread = machine__findnew_thread(machine, sample->pid, sample->tid);
2385         if (thread == NULL)
2386                 return;
2387
2388         /* show wakeup unless both awakee and awaker are filtered */
2389         if (timehist_skip_sample(sched, thread, evsel, sample) &&
2390             timehist_skip_sample(sched, awakened, evsel, sample)) {
2391                 return;
2392         }
2393
2394         timestamp__scnprintf_usec(sample->time, tstr, sizeof(tstr));
2395         printf("%15s [%04d] ", tstr, sample->cpu);
2396         if (sched->show_cpu_visual)
2397                 printf(" %*s ", sched->max_cpu.cpu + 1, "");
2398
2399         printf(" %-*s ", comm_width, timehist_get_commstr(thread));
2400
2401         /* dt spacer */
2402         printf("  %9s  %9s  %9s ", "", "", "");
2403
2404         printf("awakened: %s", timehist_get_commstr(awakened));
2405
2406         printf("\n");
2407 }
2408
2409 static int timehist_sched_wakeup_ignore(struct perf_tool *tool __maybe_unused,
2410                                         union perf_event *event __maybe_unused,
2411                                         struct evsel *evsel __maybe_unused,
2412                                         struct perf_sample *sample __maybe_unused,
2413                                         struct machine *machine __maybe_unused)
2414 {
2415         return 0;
2416 }
2417
2418 static int timehist_sched_wakeup_event(struct perf_tool *tool,
2419                                        union perf_event *event __maybe_unused,
2420                                        struct evsel *evsel,
2421                                        struct perf_sample *sample,
2422                                        struct machine *machine)
2423 {
2424         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
2425         struct thread *thread;
2426         struct thread_runtime *tr = NULL;
2427         /* want pid of awakened task not pid in sample */
2428         const u32 pid = evsel__intval(evsel, sample, "pid");
2429
2430         thread = machine__findnew_thread(machine, 0, pid);
2431         if (thread == NULL)
2432                 return -1;
2433
2434         tr = thread__get_runtime(thread);
2435         if (tr == NULL)
2436                 return -1;
2437
2438         if (tr->ready_to_run == 0)
2439                 tr->ready_to_run = sample->time;
2440
2441         /* show wakeups if requested */
2442         if (sched->show_wakeups &&
2443             !perf_time__skip_sample(&sched->ptime, sample->time))
2444                 timehist_print_wakeup_event(sched, evsel, sample, machine, thread);
2445
2446         return 0;
2447 }
2448
2449 static void timehist_print_migration_event(struct perf_sched *sched,
2450                                         struct evsel *evsel,
2451                                         struct perf_sample *sample,
2452                                         struct machine *machine,
2453                                         struct thread *migrated)
2454 {
2455         struct thread *thread;
2456         char tstr[64];
2457         u32 max_cpus;
2458         u32 ocpu, dcpu;
2459
2460         if (sched->summary_only)
2461                 return;
2462
2463         max_cpus = sched->max_cpu.cpu + 1;
2464         ocpu = evsel__intval(evsel, sample, "orig_cpu");
2465         dcpu = evsel__intval(evsel, sample, "dest_cpu");
2466
2467         thread = machine__findnew_thread(machine, sample->pid, sample->tid);
2468         if (thread == NULL)
2469                 return;
2470
2471         if (timehist_skip_sample(sched, thread, evsel, sample) &&
2472             timehist_skip_sample(sched, migrated, evsel, sample)) {
2473                 return;
2474         }
2475
2476         timestamp__scnprintf_usec(sample->time, tstr, sizeof(tstr));
2477         printf("%15s [%04d] ", tstr, sample->cpu);
2478
2479         if (sched->show_cpu_visual) {
2480                 u32 i;
2481                 char c;
2482
2483                 printf("  ");
2484                 for (i = 0; i < max_cpus; ++i) {
2485                         c = (i == sample->cpu) ? 'm' : ' ';
2486                         printf("%c", c);
2487                 }
2488                 printf("  ");
2489         }
2490
2491         printf(" %-*s ", comm_width, timehist_get_commstr(thread));
2492
2493         /* dt spacer */
2494         printf("  %9s  %9s  %9s ", "", "", "");
2495
2496         printf("migrated: %s", timehist_get_commstr(migrated));
2497         printf(" cpu %d => %d", ocpu, dcpu);
2498
2499         printf("\n");
2500 }
2501
2502 static int timehist_migrate_task_event(struct perf_tool *tool,
2503                                        union perf_event *event __maybe_unused,
2504                                        struct evsel *evsel,
2505                                        struct perf_sample *sample,
2506                                        struct machine *machine)
2507 {
2508         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
2509         struct thread *thread;
2510         struct thread_runtime *tr = NULL;
2511         /* want pid of migrated task not pid in sample */
2512         const u32 pid = evsel__intval(evsel, sample, "pid");
2513
2514         thread = machine__findnew_thread(machine, 0, pid);
2515         if (thread == NULL)
2516                 return -1;
2517
2518         tr = thread__get_runtime(thread);
2519         if (tr == NULL)
2520                 return -1;
2521
2522         tr->migrations++;
2523
2524         /* show migrations if requested */
2525         timehist_print_migration_event(sched, evsel, sample, machine, thread);
2526
2527         return 0;
2528 }
2529
2530 static int timehist_sched_change_event(struct perf_tool *tool,
2531                                        union perf_event *event,
2532                                        struct evsel *evsel,
2533                                        struct perf_sample *sample,
2534                                        struct machine *machine)
2535 {
2536         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
2537         struct perf_time_interval *ptime = &sched->ptime;
2538         struct addr_location al;
2539         struct thread *thread;
2540         struct thread_runtime *tr = NULL;
2541         u64 tprev, t = sample->time;
2542         int rc = 0;
2543         int state = evsel__intval(evsel, sample, "prev_state");
2544
2545         if (machine__resolve(machine, &al, sample) < 0) {
2546                 pr_err("problem processing %d event. skipping it\n",
2547                        event->header.type);
2548                 rc = -1;
2549                 goto out;
2550         }
2551
2552         thread = timehist_get_thread(sched, sample, machine, evsel);
2553         if (thread == NULL) {
2554                 rc = -1;
2555                 goto out;
2556         }
2557
2558         if (timehist_skip_sample(sched, thread, evsel, sample))
2559                 goto out;
2560
2561         tr = thread__get_runtime(thread);
2562         if (tr == NULL) {
2563                 rc = -1;
2564                 goto out;
2565         }
2566
2567         tprev = evsel__get_time(evsel, sample->cpu);
2568
2569         /*
2570          * If start time given:
2571          * - sample time is under window user cares about - skip sample
2572          * - tprev is under window user cares about  - reset to start of window
2573          */
2574         if (ptime->start && ptime->start > t)
2575                 goto out;
2576
2577         if (tprev && ptime->start > tprev)
2578                 tprev = ptime->start;
2579
2580         /*
2581          * If end time given:
2582          * - previous sched event is out of window - we are done
2583          * - sample time is beyond window user cares about - reset it
2584          *   to close out stats for time window interest
2585          */
2586         if (ptime->end) {
2587                 if (tprev > ptime->end)
2588                         goto out;
2589
2590                 if (t > ptime->end)
2591                         t = ptime->end;
2592         }
2593
2594         if (!sched->idle_hist || thread->tid == 0) {
2595                 if (!cpu_list || test_bit(sample->cpu, cpu_bitmap))
2596                         timehist_update_runtime_stats(tr, t, tprev);
2597
2598                 if (sched->idle_hist) {
2599                         struct idle_thread_runtime *itr = (void *)tr;
2600                         struct thread_runtime *last_tr;
2601
2602                         BUG_ON(thread->tid != 0);
2603
2604                         if (itr->last_thread == NULL)
2605                                 goto out;
2606
2607                         /* add current idle time as last thread's runtime */
2608                         last_tr = thread__get_runtime(itr->last_thread);
2609                         if (last_tr == NULL)
2610                                 goto out;
2611
2612                         timehist_update_runtime_stats(last_tr, t, tprev);
2613                         /*
2614                          * remove delta time of last thread as it's not updated
2615                          * and otherwise it will show an invalid value next
2616                          * time.  we only care total run time and run stat.
2617                          */
2618                         last_tr->dt_run = 0;
2619                         last_tr->dt_delay = 0;
2620                         last_tr->dt_sleep = 0;
2621                         last_tr->dt_iowait = 0;
2622                         last_tr->dt_preempt = 0;
2623
2624                         if (itr->cursor.nr)
2625                                 callchain_append(&itr->callchain, &itr->cursor, t - tprev);
2626
2627                         itr->last_thread = NULL;
2628                 }
2629         }
2630
2631         if (!sched->summary_only)
2632                 timehist_print_sample(sched, evsel, sample, &al, thread, t, state);
2633
2634 out:
2635         if (sched->hist_time.start == 0 && t >= ptime->start)
2636                 sched->hist_time.start = t;
2637         if (ptime->end == 0 || t <= ptime->end)
2638                 sched->hist_time.end = t;
2639
2640         if (tr) {
2641                 /* time of this sched_switch event becomes last time task seen */
2642                 tr->last_time = sample->time;
2643
2644                 /* last state is used to determine where to account wait time */
2645                 tr->last_state = state;
2646
2647                 /* sched out event for task so reset ready to run time */
2648                 tr->ready_to_run = 0;
2649         }
2650
2651         evsel__save_time(evsel, sample->time, sample->cpu);
2652
2653         return rc;
2654 }
2655
2656 static int timehist_sched_switch_event(struct perf_tool *tool,
2657                              union perf_event *event,
2658                              struct evsel *evsel,
2659                              struct perf_sample *sample,
2660                              struct machine *machine __maybe_unused)
2661 {
2662         return timehist_sched_change_event(tool, event, evsel, sample, machine);
2663 }
2664
2665 static int process_lost(struct perf_tool *tool __maybe_unused,
2666                         union perf_event *event,
2667                         struct perf_sample *sample,
2668                         struct machine *machine __maybe_unused)
2669 {
2670         char tstr[64];
2671
2672         timestamp__scnprintf_usec(sample->time, tstr, sizeof(tstr));
2673         printf("%15s ", tstr);
2674         printf("lost %" PRI_lu64 " events on cpu %d\n", event->lost.lost, sample->cpu);
2675
2676         return 0;
2677 }
2678
2679
2680 static void print_thread_runtime(struct thread *t,
2681                                  struct thread_runtime *r)
2682 {
2683         double mean = avg_stats(&r->run_stats);
2684         float stddev;
2685
2686         printf("%*s   %5d  %9" PRIu64 " ",
2687                comm_width, timehist_get_commstr(t), t->ppid,
2688                (u64) r->run_stats.n);
2689
2690         print_sched_time(r->total_run_time, 8);
2691         stddev = rel_stddev_stats(stddev_stats(&r->run_stats), mean);
2692         print_sched_time(r->run_stats.min, 6);
2693         printf(" ");
2694         print_sched_time((u64) mean, 6);
2695         printf(" ");
2696         print_sched_time(r->run_stats.max, 6);
2697         printf("  ");
2698         printf("%5.2f", stddev);
2699         printf("   %5" PRIu64, r->migrations);
2700         printf("\n");
2701 }
2702
2703 static void print_thread_waittime(struct thread *t,
2704                                   struct thread_runtime *r)
2705 {
2706         printf("%*s   %5d  %9" PRIu64 " ",
2707                comm_width, timehist_get_commstr(t), t->ppid,
2708                (u64) r->run_stats.n);
2709
2710         print_sched_time(r->total_run_time, 8);
2711         print_sched_time(r->total_sleep_time, 6);
2712         printf(" ");
2713         print_sched_time(r->total_iowait_time, 6);
2714         printf(" ");
2715         print_sched_time(r->total_preempt_time, 6);
2716         printf(" ");
2717         print_sched_time(r->total_delay_time, 6);
2718         printf("\n");
2719 }
2720
2721 struct total_run_stats {
2722         struct perf_sched *sched;
2723         u64  sched_count;
2724         u64  task_count;
2725         u64  total_run_time;
2726 };
2727
2728 static int __show_thread_runtime(struct thread *t, void *priv)
2729 {
2730         struct total_run_stats *stats = priv;
2731         struct thread_runtime *r;
2732
2733         if (thread__is_filtered(t))
2734                 return 0;
2735
2736         r = thread__priv(t);
2737         if (r && r->run_stats.n) {
2738                 stats->task_count++;
2739                 stats->sched_count += r->run_stats.n;
2740                 stats->total_run_time += r->total_run_time;
2741
2742                 if (stats->sched->show_state)
2743                         print_thread_waittime(t, r);
2744                 else
2745                         print_thread_runtime(t, r);
2746         }
2747
2748         return 0;
2749 }
2750
2751 static int show_thread_runtime(struct thread *t, void *priv)
2752 {
2753         if (t->dead)
2754                 return 0;
2755
2756         return __show_thread_runtime(t, priv);
2757 }
2758
2759 static int show_deadthread_runtime(struct thread *t, void *priv)
2760 {
2761         if (!t->dead)
2762                 return 0;
2763
2764         return __show_thread_runtime(t, priv);
2765 }
2766
2767 static size_t callchain__fprintf_folded(FILE *fp, struct callchain_node *node)
2768 {
2769         const char *sep = " <- ";
2770         struct callchain_list *chain;
2771         size_t ret = 0;
2772         char bf[1024];
2773         bool first;
2774
2775         if (node == NULL)
2776                 return 0;
2777
2778         ret = callchain__fprintf_folded(fp, node->parent);
2779         first = (ret == 0);
2780
2781         list_for_each_entry(chain, &node->val, list) {
2782                 if (chain->ip >= PERF_CONTEXT_MAX)
2783                         continue;
2784                 if (chain->ms.sym && chain->ms.sym->ignore)
2785                         continue;
2786                 ret += fprintf(fp, "%s%s", first ? "" : sep,
2787                                callchain_list__sym_name(chain, bf, sizeof(bf),
2788                                                         false));
2789                 first = false;
2790         }
2791
2792         return ret;
2793 }
2794
2795 static size_t timehist_print_idlehist_callchain(struct rb_root_cached *root)
2796 {
2797         size_t ret = 0;
2798         FILE *fp = stdout;
2799         struct callchain_node *chain;
2800         struct rb_node *rb_node = rb_first_cached(root);
2801
2802         printf("  %16s  %8s  %s\n", "Idle time (msec)", "Count", "Callchains");
2803         printf("  %.16s  %.8s  %.50s\n", graph_dotted_line, graph_dotted_line,
2804                graph_dotted_line);
2805
2806         while (rb_node) {
2807                 chain = rb_entry(rb_node, struct callchain_node, rb_node);
2808                 rb_node = rb_next(rb_node);
2809
2810                 ret += fprintf(fp, "  ");
2811                 print_sched_time(chain->hit, 12);
2812                 ret += 16;  /* print_sched_time returns 2nd arg + 4 */
2813                 ret += fprintf(fp, " %8d  ", chain->count);
2814                 ret += callchain__fprintf_folded(fp, chain);
2815                 ret += fprintf(fp, "\n");
2816         }
2817
2818         return ret;
2819 }
2820
2821 static void timehist_print_summary(struct perf_sched *sched,
2822                                    struct perf_session *session)
2823 {
2824         struct machine *m = &session->machines.host;
2825         struct total_run_stats totals;
2826         u64 task_count;
2827         struct thread *t;
2828         struct thread_runtime *r;
2829         int i;
2830         u64 hist_time = sched->hist_time.end - sched->hist_time.start;
2831
2832         memset(&totals, 0, sizeof(totals));
2833         totals.sched = sched;
2834
2835         if (sched->idle_hist) {
2836                 printf("\nIdle-time summary\n");
2837                 printf("%*s  parent  sched-out  ", comm_width, "comm");
2838                 printf("  idle-time   min-idle    avg-idle    max-idle  stddev  migrations\n");
2839         } else if (sched->show_state) {
2840                 printf("\nWait-time summary\n");
2841                 printf("%*s  parent   sched-in  ", comm_width, "comm");
2842                 printf("   run-time      sleep      iowait     preempt       delay\n");
2843         } else {
2844                 printf("\nRuntime summary\n");
2845                 printf("%*s  parent   sched-in  ", comm_width, "comm");
2846                 printf("   run-time    min-run     avg-run     max-run  stddev  migrations\n");
2847         }
2848         printf("%*s            (count)  ", comm_width, "");
2849         printf("     (msec)     (msec)      (msec)      (msec)       %s\n",
2850                sched->show_state ? "(msec)" : "%");
2851         printf("%.117s\n", graph_dotted_line);
2852
2853         machine__for_each_thread(m, show_thread_runtime, &totals);
2854         task_count = totals.task_count;
2855         if (!task_count)
2856                 printf("<no still running tasks>\n");
2857
2858         printf("\nTerminated tasks:\n");
2859         machine__for_each_thread(m, show_deadthread_runtime, &totals);
2860         if (task_count == totals.task_count)
2861                 printf("<no terminated tasks>\n");
2862
2863         /* CPU idle stats not tracked when samples were skipped */
2864         if (sched->skipped_samples && !sched->idle_hist)
2865                 return;
2866
2867         printf("\nIdle stats:\n");
2868         for (i = 0; i < idle_max_cpu; ++i) {
2869                 if (cpu_list && !test_bit(i, cpu_bitmap))
2870                         continue;
2871
2872                 t = idle_threads[i];
2873                 if (!t)
2874                         continue;
2875
2876                 r = thread__priv(t);
2877                 if (r && r->run_stats.n) {
2878                         totals.sched_count += r->run_stats.n;
2879                         printf("    CPU %2d idle for ", i);
2880                         print_sched_time(r->total_run_time, 6);
2881                         printf(" msec  (%6.2f%%)\n", 100.0 * r->total_run_time / hist_time);
2882                 } else
2883                         printf("    CPU %2d idle entire time window\n", i);
2884         }
2885
2886         if (sched->idle_hist && sched->show_callchain) {
2887                 callchain_param.mode  = CHAIN_FOLDED;
2888                 callchain_param.value = CCVAL_PERIOD;
2889
2890                 callchain_register_param(&callchain_param);
2891
2892                 printf("\nIdle stats by callchain:\n");
2893                 for (i = 0; i < idle_max_cpu; ++i) {
2894                         struct idle_thread_runtime *itr;
2895
2896                         t = idle_threads[i];
2897                         if (!t)
2898                                 continue;
2899
2900                         itr = thread__priv(t);
2901                         if (itr == NULL)
2902                                 continue;
2903
2904                         callchain_param.sort(&itr->sorted_root.rb_root, &itr->callchain,
2905                                              0, &callchain_param);
2906
2907                         printf("  CPU %2d:", i);
2908                         print_sched_time(itr->tr.total_run_time, 6);
2909                         printf(" msec\n");
2910                         timehist_print_idlehist_callchain(&itr->sorted_root);
2911                         printf("\n");
2912                 }
2913         }
2914
2915         printf("\n"
2916                "    Total number of unique tasks: %" PRIu64 "\n"
2917                "Total number of context switches: %" PRIu64 "\n",
2918                totals.task_count, totals.sched_count);
2919
2920         printf("           Total run time (msec): ");
2921         print_sched_time(totals.total_run_time, 2);
2922         printf("\n");
2923
2924         printf("    Total scheduling time (msec): ");
2925         print_sched_time(hist_time, 2);
2926         printf(" (x %d)\n", sched->max_cpu.cpu);
2927 }
2928
2929 typedef int (*sched_handler)(struct perf_tool *tool,
2930                           union perf_event *event,
2931                           struct evsel *evsel,
2932                           struct perf_sample *sample,
2933                           struct machine *machine);
2934
2935 static int perf_timehist__process_sample(struct perf_tool *tool,
2936                                          union perf_event *event,
2937                                          struct perf_sample *sample,
2938                                          struct evsel *evsel,
2939                                          struct machine *machine)
2940 {
2941         struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
2942         int err = 0;
2943         struct perf_cpu this_cpu = {
2944                 .cpu = sample->cpu,
2945         };
2946
2947         if (this_cpu.cpu > sched->max_cpu.cpu)
2948                 sched->max_cpu = this_cpu;
2949
2950         if (evsel->handler != NULL) {
2951                 sched_handler f = evsel->handler;
2952
2953                 err = f(tool, event, evsel, sample, machine);
2954         }
2955
2956         return err;
2957 }
2958
2959 static int timehist_check_attr(struct perf_sched *sched,
2960                                struct evlist *evlist)
2961 {
2962         struct evsel *evsel;
2963         struct evsel_runtime *er;
2964
2965         list_for_each_entry(evsel, &evlist->core.entries, core.node) {
2966                 er = evsel__get_runtime(evsel);
2967                 if (er == NULL) {
2968                         pr_err("Failed to allocate memory for evsel runtime data\n");
2969                         return -1;
2970                 }
2971
2972                 if (sched->show_callchain && !evsel__has_callchain(evsel)) {
2973                         pr_info("Samples do not have callchains.\n");
2974                         sched->show_callchain = 0;
2975                         symbol_conf.use_callchain = 0;
2976                 }
2977         }
2978
2979         return 0;
2980 }
2981
2982 static int perf_sched__timehist(struct perf_sched *sched)
2983 {
2984         struct evsel_str_handler handlers[] = {
2985                 { "sched:sched_switch",       timehist_sched_switch_event, },
2986                 { "sched:sched_wakeup",       timehist_sched_wakeup_event, },
2987                 { "sched:sched_waking",       timehist_sched_wakeup_event, },
2988                 { "sched:sched_wakeup_new",   timehist_sched_wakeup_event, },
2989         };
2990         const struct evsel_str_handler migrate_handlers[] = {
2991                 { "sched:sched_migrate_task", timehist_migrate_task_event, },
2992         };
2993         struct perf_data data = {
2994                 .path  = input_name,
2995                 .mode  = PERF_DATA_MODE_READ,
2996                 .force = sched->force,
2997         };
2998
2999         struct perf_session *session;
3000         struct evlist *evlist;
3001         int err = -1;
3002
3003         /*
3004          * event handlers for timehist option
3005          */
3006         sched->tool.sample       = perf_timehist__process_sample;
3007         sched->tool.mmap         = perf_event__process_mmap;
3008         sched->tool.comm         = perf_event__process_comm;
3009         sched->tool.exit         = perf_event__process_exit;
3010         sched->tool.fork         = perf_event__process_fork;
3011         sched->tool.lost         = process_lost;
3012         sched->tool.attr         = perf_event__process_attr;
3013         sched->tool.tracing_data = perf_event__process_tracing_data;
3014         sched->tool.build_id     = perf_event__process_build_id;
3015
3016         sched->tool.ordered_events = true;
3017         sched->tool.ordering_requires_timestamps = true;
3018
3019         symbol_conf.use_callchain = sched->show_callchain;
3020
3021         session = perf_session__new(&data, &sched->tool);
3022         if (IS_ERR(session))
3023                 return PTR_ERR(session);
3024
3025         if (cpu_list) {
3026                 err = perf_session__cpu_bitmap(session, cpu_list, cpu_bitmap);
3027                 if (err < 0)
3028                         goto out;
3029         }
3030
3031         evlist = session->evlist;
3032
3033         symbol__init(&session->header.env);
3034
3035         if (perf_time__parse_str(&sched->ptime, sched->time_str) != 0) {
3036                 pr_err("Invalid time string\n");
3037                 return -EINVAL;
3038         }
3039
3040         if (timehist_check_attr(sched, evlist) != 0)
3041                 goto out;
3042
3043         setup_pager();
3044
3045         /* prefer sched_waking if it is captured */
3046         if (evlist__find_tracepoint_by_name(session->evlist, "sched:sched_waking"))
3047                 handlers[1].handler = timehist_sched_wakeup_ignore;
3048
3049         /* setup per-evsel handlers */
3050         if (perf_session__set_tracepoints_handlers(session, handlers))
3051                 goto out;
3052
3053         /* sched_switch event at a minimum needs to exist */
3054         if (!evlist__find_tracepoint_by_name(session->evlist, "sched:sched_switch")) {
3055                 pr_err("No sched_switch events found. Have you run 'perf sched record'?\n");
3056                 goto out;
3057         }
3058
3059         if (sched->show_migrations &&
3060             perf_session__set_tracepoints_handlers(session, migrate_handlers))
3061                 goto out;
3062
3063         /* pre-allocate struct for per-CPU idle stats */
3064         sched->max_cpu.cpu = session->header.env.nr_cpus_online;
3065         if (sched->max_cpu.cpu == 0)
3066                 sched->max_cpu.cpu = 4;
3067         if (init_idle_threads(sched->max_cpu.cpu))
3068                 goto out;
3069
3070         /* summary_only implies summary option, but don't overwrite summary if set */
3071         if (sched->summary_only)
3072                 sched->summary = sched->summary_only;
3073
3074         if (!sched->summary_only)
3075                 timehist_header(sched);
3076
3077         err = perf_session__process_events(session);
3078         if (err) {
3079                 pr_err("Failed to process events, error %d", err);
3080                 goto out;
3081         }
3082
3083         sched->nr_events      = evlist->stats.nr_events[0];
3084         sched->nr_lost_events = evlist->stats.total_lost;
3085         sched->nr_lost_chunks = evlist->stats.nr_events[PERF_RECORD_LOST];
3086
3087         if (sched->summary)
3088                 timehist_print_summary(sched, session);
3089
3090 out:
3091         free_idle_threads();
3092         perf_session__delete(session);
3093
3094         return err;
3095 }
3096
3097
3098 static void print_bad_events(struct perf_sched *sched)
3099 {
3100         if (sched->nr_unordered_timestamps && sched->nr_timestamps) {
3101                 printf("  INFO: %.3f%% unordered timestamps (%ld out of %ld)\n",
3102                         (double)sched->nr_unordered_timestamps/(double)sched->nr_timestamps*100.0,
3103                         sched->nr_unordered_timestamps, sched->nr_timestamps);
3104         }
3105         if (sched->nr_lost_events && sched->nr_events) {
3106                 printf("  INFO: %.3f%% lost events (%ld out of %ld, in %ld chunks)\n",
3107                         (double)sched->nr_lost_events/(double)sched->nr_events * 100.0,
3108                         sched->nr_lost_events, sched->nr_events, sched->nr_lost_chunks);
3109         }
3110         if (sched->nr_context_switch_bugs && sched->nr_timestamps) {
3111                 printf("  INFO: %.3f%% context switch bugs (%ld out of %ld)",
3112                         (double)sched->nr_context_switch_bugs/(double)sched->nr_timestamps*100.0,
3113                         sched->nr_context_switch_bugs, sched->nr_timestamps);
3114                 if (sched->nr_lost_events)
3115                         printf(" (due to lost events?)");
3116                 printf("\n");
3117         }
3118 }
3119
3120 static void __merge_work_atoms(struct rb_root_cached *root, struct work_atoms *data)
3121 {
3122         struct rb_node **new = &(root->rb_root.rb_node), *parent = NULL;
3123         struct work_atoms *this;
3124         const char *comm = thread__comm_str(data->thread), *this_comm;
3125         bool leftmost = true;
3126
3127         while (*new) {
3128                 int cmp;
3129
3130                 this = container_of(*new, struct work_atoms, node);
3131                 parent = *new;
3132
3133                 this_comm = thread__comm_str(this->thread);
3134                 cmp = strcmp(comm, this_comm);
3135                 if (cmp > 0) {
3136                         new = &((*new)->rb_left);
3137                 } else if (cmp < 0) {
3138                         new = &((*new)->rb_right);
3139                         leftmost = false;
3140                 } else {
3141                         this->num_merged++;
3142                         this->total_runtime += data->total_runtime;
3143                         this->nb_atoms += data->nb_atoms;
3144                         this->total_lat += data->total_lat;
3145                         list_splice(&data->work_list, &this->work_list);
3146                         if (this->max_lat < data->max_lat) {
3147                                 this->max_lat = data->max_lat;
3148                                 this->max_lat_start = data->max_lat_start;
3149                                 this->max_lat_end = data->max_lat_end;
3150                         }
3151                         zfree(&data);
3152                         return;
3153                 }
3154         }
3155
3156         data->num_merged++;
3157         rb_link_node(&data->node, parent, new);
3158         rb_insert_color_cached(&data->node, root, leftmost);
3159 }
3160
3161 static void perf_sched__merge_lat(struct perf_sched *sched)
3162 {
3163         struct work_atoms *data;
3164         struct rb_node *node;
3165
3166         if (sched->skip_merge)
3167                 return;
3168
3169         while ((node = rb_first_cached(&sched->atom_root))) {
3170                 rb_erase_cached(node, &sched->atom_root);
3171                 data = rb_entry(node, struct work_atoms, node);
3172                 __merge_work_atoms(&sched->merged_atom_root, data);
3173         }
3174 }
3175
3176 static int perf_sched__lat(struct perf_sched *sched)
3177 {
3178         struct rb_node *next;
3179
3180         setup_pager();
3181
3182         if (perf_sched__read_events(sched))
3183                 return -1;
3184
3185         perf_sched__merge_lat(sched);
3186         perf_sched__sort_lat(sched);
3187
3188         printf("\n -------------------------------------------------------------------------------------------------------------------------------------------\n");
3189         printf("  Task                  |   Runtime ms  | Switches | Avg delay ms    | Max delay ms    | Max delay start           | Max delay end          |\n");
3190         printf(" -------------------------------------------------------------------------------------------------------------------------------------------\n");
3191
3192         next = rb_first_cached(&sched->sorted_atom_root);
3193
3194         while (next) {
3195                 struct work_atoms *work_list;
3196
3197                 work_list = rb_entry(next, struct work_atoms, node);
3198                 output_lat_thread(sched, work_list);
3199                 next = rb_next(next);
3200                 thread__zput(work_list->thread);
3201         }
3202
3203         printf(" -----------------------------------------------------------------------------------------------------------------\n");
3204         printf("  TOTAL:                |%11.3f ms |%9" PRIu64 " |\n",
3205                 (double)sched->all_runtime / NSEC_PER_MSEC, sched->all_count);
3206
3207         printf(" ---------------------------------------------------\n");
3208
3209         print_bad_events(sched);
3210         printf("\n");
3211
3212         return 0;
3213 }
3214
3215 static int setup_map_cpus(struct perf_sched *sched)
3216 {
3217         struct perf_cpu_map *map;
3218
3219         sched->max_cpu.cpu  = sysconf(_SC_NPROCESSORS_CONF);
3220
3221         if (sched->map.comp) {
3222                 sched->map.comp_cpus = zalloc(sched->max_cpu.cpu * sizeof(int));
3223                 if (!sched->map.comp_cpus)
3224                         return -1;
3225         }
3226
3227         if (!sched->map.cpus_str)
3228                 return 0;
3229
3230         map = perf_cpu_map__new(sched->map.cpus_str);
3231         if (!map) {
3232                 pr_err("failed to get cpus map from %s\n", sched->map.cpus_str);
3233                 return -1;
3234         }
3235
3236         sched->map.cpus = map;
3237         return 0;
3238 }
3239
3240 static int setup_color_pids(struct perf_sched *sched)
3241 {
3242         struct perf_thread_map *map;
3243
3244         if (!sched->map.color_pids_str)
3245                 return 0;
3246
3247         map = thread_map__new_by_tid_str(sched->map.color_pids_str);
3248         if (!map) {
3249                 pr_err("failed to get thread map from %s\n", sched->map.color_pids_str);
3250                 return -1;
3251         }
3252
3253         sched->map.color_pids = map;
3254         return 0;
3255 }
3256
3257 static int setup_color_cpus(struct perf_sched *sched)
3258 {
3259         struct perf_cpu_map *map;
3260
3261         if (!sched->map.color_cpus_str)
3262                 return 0;
3263
3264         map = perf_cpu_map__new(sched->map.color_cpus_str);
3265         if (!map) {
3266                 pr_err("failed to get thread map from %s\n", sched->map.color_cpus_str);
3267                 return -1;
3268         }
3269
3270         sched->map.color_cpus = map;
3271         return 0;
3272 }
3273
3274 static int perf_sched__map(struct perf_sched *sched)
3275 {
3276         if (setup_map_cpus(sched))
3277                 return -1;
3278
3279         if (setup_color_pids(sched))
3280                 return -1;
3281
3282         if (setup_color_cpus(sched))
3283                 return -1;
3284
3285         setup_pager();
3286         if (perf_sched__read_events(sched))
3287                 return -1;
3288         print_bad_events(sched);
3289         return 0;
3290 }
3291
3292 static int perf_sched__replay(struct perf_sched *sched)
3293 {
3294         unsigned long i;
3295
3296         calibrate_run_measurement_overhead(sched);
3297         calibrate_sleep_measurement_overhead(sched);
3298
3299         test_calibrations(sched);
3300
3301         if (perf_sched__read_events(sched))
3302                 return -1;
3303
3304         printf("nr_run_events:        %ld\n", sched->nr_run_events);
3305         printf("nr_sleep_events:      %ld\n", sched->nr_sleep_events);
3306         printf("nr_wakeup_events:     %ld\n", sched->nr_wakeup_events);
3307
3308         if (sched->targetless_wakeups)
3309                 printf("target-less wakeups:  %ld\n", sched->targetless_wakeups);
3310         if (sched->multitarget_wakeups)
3311                 printf("multi-target wakeups: %ld\n", sched->multitarget_wakeups);
3312         if (sched->nr_run_events_optimized)
3313                 printf("run atoms optimized: %ld\n",
3314                         sched->nr_run_events_optimized);
3315
3316         print_task_traces(sched);
3317         add_cross_task_wakeups(sched);
3318
3319         create_tasks(sched);
3320         printf("------------------------------------------------------------\n");
3321         for (i = 0; i < sched->replay_repeat; i++)
3322                 run_one_test(sched);
3323
3324         return 0;
3325 }
3326
3327 static void setup_sorting(struct perf_sched *sched, const struct option *options,
3328                           const char * const usage_msg[])
3329 {
3330         char *tmp, *tok, *str = strdup(sched->sort_order);
3331
3332         for (tok = strtok_r(str, ", ", &tmp);
3333                         tok; tok = strtok_r(NULL, ", ", &tmp)) {
3334                 if (sort_dimension__add(tok, &sched->sort_list) < 0) {
3335                         usage_with_options_msg(usage_msg, options,
3336                                         "Unknown --sort key: `%s'", tok);
3337                 }
3338         }
3339
3340         free(str);
3341
3342         sort_dimension__add("pid", &sched->cmp_pid);
3343 }
3344
3345 static bool schedstat_events_exposed(void)
3346 {
3347         /*
3348          * Select "sched:sched_stat_wait" event to check
3349          * whether schedstat tracepoints are exposed.
3350          */
3351         return IS_ERR(trace_event__tp_format("sched", "sched_stat_wait")) ?
3352                 false : true;
3353 }
3354
3355 static int __cmd_record(int argc, const char **argv)
3356 {
3357         unsigned int rec_argc, i, j;
3358         const char **rec_argv;
3359         const char * const record_args[] = {
3360                 "record",
3361                 "-a",
3362                 "-R",
3363                 "-m", "1024",
3364                 "-c", "1",
3365                 "-e", "sched:sched_switch",
3366                 "-e", "sched:sched_stat_runtime",
3367                 "-e", "sched:sched_process_fork",
3368                 "-e", "sched:sched_wakeup_new",
3369                 "-e", "sched:sched_migrate_task",
3370         };
3371
3372         /*
3373          * The tracepoints trace_sched_stat_{wait, sleep, iowait}
3374          * are not exposed to user if CONFIG_SCHEDSTATS is not set,
3375          * to prevent "perf sched record" execution failure, determine
3376          * whether to record schedstat events according to actual situation.
3377          */
3378         const char * const schedstat_args[] = {
3379                 "-e", "sched:sched_stat_wait",
3380                 "-e", "sched:sched_stat_sleep",
3381                 "-e", "sched:sched_stat_iowait",
3382         };
3383         unsigned int schedstat_argc = schedstat_events_exposed() ?
3384                 ARRAY_SIZE(schedstat_args) : 0;
3385
3386         struct tep_event *waking_event;
3387
3388         /*
3389          * +2 for either "-e", "sched:sched_wakeup" or
3390          * "-e", "sched:sched_waking"
3391          */
3392         rec_argc = ARRAY_SIZE(record_args) + 2 + schedstat_argc + argc - 1;
3393         rec_argv = calloc(rec_argc + 1, sizeof(char *));
3394
3395         if (rec_argv == NULL)
3396                 return -ENOMEM;
3397
3398         for (i = 0; i < ARRAY_SIZE(record_args); i++)
3399                 rec_argv[i] = strdup(record_args[i]);
3400
3401         rec_argv[i++] = "-e";
3402         waking_event = trace_event__tp_format("sched", "sched_waking");
3403         if (!IS_ERR(waking_event))
3404                 rec_argv[i++] = strdup("sched:sched_waking");
3405         else
3406                 rec_argv[i++] = strdup("sched:sched_wakeup");
3407
3408         for (j = 0; j < schedstat_argc; j++)
3409                 rec_argv[i++] = strdup(schedstat_args[j]);
3410
3411         for (j = 1; j < (unsigned int)argc; j++, i++)
3412                 rec_argv[i] = argv[j];
3413
3414         BUG_ON(i != rec_argc);
3415
3416         return cmd_record(i, rec_argv);
3417 }
3418
3419 int cmd_sched(int argc, const char **argv)
3420 {
3421         static const char default_sort_order[] = "avg, max, switch, runtime";
3422         struct perf_sched sched = {
3423                 .tool = {
3424                         .sample          = perf_sched__process_tracepoint_sample,
3425                         .comm            = perf_sched__process_comm,
3426                         .namespaces      = perf_event__process_namespaces,
3427                         .lost            = perf_event__process_lost,
3428                         .fork            = perf_sched__process_fork_event,
3429                         .ordered_events = true,
3430                 },
3431                 .cmp_pid              = LIST_HEAD_INIT(sched.cmp_pid),
3432                 .sort_list            = LIST_HEAD_INIT(sched.sort_list),
3433                 .start_work_mutex     = PTHREAD_MUTEX_INITIALIZER,
3434                 .work_done_wait_mutex = PTHREAD_MUTEX_INITIALIZER,
3435                 .sort_order           = default_sort_order,
3436                 .replay_repeat        = 10,
3437                 .profile_cpu          = -1,
3438                 .next_shortname1      = 'A',
3439                 .next_shortname2      = '0',
3440                 .skip_merge           = 0,
3441                 .show_callchain       = 1,
3442                 .max_stack            = 5,
3443         };
3444         const struct option sched_options[] = {
3445         OPT_STRING('i', "input", &input_name, "file",
3446                     "input file name"),
3447         OPT_INCR('v', "verbose", &verbose,
3448                     "be more verbose (show symbol address, etc)"),
3449         OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
3450                     "dump raw trace in ASCII"),
3451         OPT_BOOLEAN('f', "force", &sched.force, "don't complain, do it"),
3452         OPT_END()
3453         };
3454         const struct option latency_options[] = {
3455         OPT_STRING('s', "sort", &sched.sort_order, "key[,key2...]",
3456                    "sort by key(s): runtime, switch, avg, max"),
3457         OPT_INTEGER('C', "CPU", &sched.profile_cpu,
3458                     "CPU to profile on"),
3459         OPT_BOOLEAN('p', "pids", &sched.skip_merge,
3460                     "latency stats per pid instead of per comm"),
3461         OPT_PARENT(sched_options)
3462         };
3463         const struct option replay_options[] = {
3464         OPT_UINTEGER('r', "repeat", &sched.replay_repeat,
3465                      "repeat the workload replay N times (-1: infinite)"),
3466         OPT_PARENT(sched_options)
3467         };
3468         const struct option map_options[] = {
3469         OPT_BOOLEAN(0, "compact", &sched.map.comp,
3470                     "map output in compact mode"),
3471         OPT_STRING(0, "color-pids", &sched.map.color_pids_str, "pids",
3472                    "highlight given pids in map"),
3473         OPT_STRING(0, "color-cpus", &sched.map.color_cpus_str, "cpus",
3474                     "highlight given CPUs in map"),
3475         OPT_STRING(0, "cpus", &sched.map.cpus_str, "cpus",
3476                     "display given CPUs in map"),
3477         OPT_PARENT(sched_options)
3478         };
3479         const struct option timehist_options[] = {
3480         OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
3481                    "file", "vmlinux pathname"),
3482         OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name,
3483                    "file", "kallsyms pathname"),
3484         OPT_BOOLEAN('g', "call-graph", &sched.show_callchain,
3485                     "Display call chains if present (default on)"),
3486         OPT_UINTEGER(0, "max-stack", &sched.max_stack,
3487                    "Maximum number of functions to display backtrace."),
3488         OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory",
3489                     "Look for files with symbols relative to this directory"),
3490         OPT_BOOLEAN('s', "summary", &sched.summary_only,
3491                     "Show only syscall summary with statistics"),
3492         OPT_BOOLEAN('S', "with-summary", &sched.summary,
3493                     "Show all syscalls and summary with statistics"),
3494         OPT_BOOLEAN('w', "wakeups", &sched.show_wakeups, "Show wakeup events"),
3495         OPT_BOOLEAN('n', "next", &sched.show_next, "Show next task"),
3496         OPT_BOOLEAN('M', "migrations", &sched.show_migrations, "Show migration events"),
3497         OPT_BOOLEAN('V', "cpu-visual", &sched.show_cpu_visual, "Add CPU visual"),
3498         OPT_BOOLEAN('I', "idle-hist", &sched.idle_hist, "Show idle events only"),
3499         OPT_STRING(0, "time", &sched.time_str, "str",
3500                    "Time span for analysis (start,stop)"),
3501         OPT_BOOLEAN(0, "state", &sched.show_state, "Show task state when sched-out"),
3502         OPT_STRING('p', "pid", &symbol_conf.pid_list_str, "pid[,pid...]",
3503                    "analyze events only for given process id(s)"),
3504         OPT_STRING('t', "tid", &symbol_conf.tid_list_str, "tid[,tid...]",
3505                    "analyze events only for given thread id(s)"),
3506         OPT_STRING('C', "cpu", &cpu_list, "cpu", "list of cpus to profile"),
3507         OPT_PARENT(sched_options)
3508         };
3509
3510         const char * const latency_usage[] = {
3511                 "perf sched latency [<options>]",
3512                 NULL
3513         };
3514         const char * const replay_usage[] = {
3515                 "perf sched replay [<options>]",
3516                 NULL
3517         };
3518         const char * const map_usage[] = {
3519                 "perf sched map [<options>]",
3520                 NULL
3521         };
3522         const char * const timehist_usage[] = {
3523                 "perf sched timehist [<options>]",
3524                 NULL
3525         };
3526         const char *const sched_subcommands[] = { "record", "latency", "map",
3527                                                   "replay", "script",
3528                                                   "timehist", NULL };
3529         const char *sched_usage[] = {
3530                 NULL,
3531                 NULL
3532         };
3533         struct trace_sched_handler lat_ops  = {
3534                 .wakeup_event       = latency_wakeup_event,
3535                 .switch_event       = latency_switch_event,
3536                 .runtime_event      = latency_runtime_event,
3537                 .migrate_task_event = latency_migrate_task_event,
3538         };
3539         struct trace_sched_handler map_ops  = {
3540                 .switch_event       = map_switch_event,
3541         };
3542         struct trace_sched_handler replay_ops  = {
3543                 .wakeup_event       = replay_wakeup_event,
3544                 .switch_event       = replay_switch_event,
3545                 .fork_event         = replay_fork_event,
3546         };
3547         unsigned int i;
3548         int ret;
3549
3550         for (i = 0; i < ARRAY_SIZE(sched.curr_pid); i++)
3551                 sched.curr_pid[i] = -1;
3552
3553         argc = parse_options_subcommand(argc, argv, sched_options, sched_subcommands,
3554                                         sched_usage, PARSE_OPT_STOP_AT_NON_OPTION);
3555         if (!argc)
3556                 usage_with_options(sched_usage, sched_options);
3557
3558         /*
3559          * Aliased to 'perf script' for now:
3560          */
3561         if (!strcmp(argv[0], "script"))
3562                 return cmd_script(argc, argv);
3563
3564         if (strlen(argv[0]) > 2 && strstarts("record", argv[0])) {
3565                 return __cmd_record(argc, argv);
3566         } else if (!strncmp(argv[0], "lat", 3)) {
3567                 sched.tp_handler = &lat_ops;
3568                 if (argc > 1) {
3569                         argc = parse_options(argc, argv, latency_options, latency_usage, 0);
3570                         if (argc)
3571                                 usage_with_options(latency_usage, latency_options);
3572                 }
3573                 setup_sorting(&sched, latency_options, latency_usage);
3574                 return perf_sched__lat(&sched);
3575         } else if (!strcmp(argv[0], "map")) {
3576                 if (argc) {
3577                         argc = parse_options(argc, argv, map_options, map_usage, 0);
3578                         if (argc)
3579                                 usage_with_options(map_usage, map_options);
3580                 }
3581                 sched.tp_handler = &map_ops;
3582                 setup_sorting(&sched, latency_options, latency_usage);
3583                 return perf_sched__map(&sched);
3584         } else if (strlen(argv[0]) > 2 && strstarts("replay", argv[0])) {
3585                 sched.tp_handler = &replay_ops;
3586                 if (argc) {
3587                         argc = parse_options(argc, argv, replay_options, replay_usage, 0);
3588                         if (argc)
3589                                 usage_with_options(replay_usage, replay_options);
3590                 }
3591                 return perf_sched__replay(&sched);
3592         } else if (!strcmp(argv[0], "timehist")) {
3593                 if (argc) {
3594                         argc = parse_options(argc, argv, timehist_options,
3595                                              timehist_usage, 0);
3596                         if (argc)
3597                                 usage_with_options(timehist_usage, timehist_options);
3598                 }
3599                 if ((sched.show_wakeups || sched.show_next) &&
3600                     sched.summary_only) {
3601                         pr_err(" Error: -s and -[n|w] are mutually exclusive.\n");
3602                         parse_options_usage(timehist_usage, timehist_options, "s", true);
3603                         if (sched.show_wakeups)
3604                                 parse_options_usage(NULL, timehist_options, "w", true);
3605                         if (sched.show_next)
3606                                 parse_options_usage(NULL, timehist_options, "n", true);
3607                         return -EINVAL;
3608                 }
3609                 ret = symbol__validate_sym_arguments();
3610                 if (ret)
3611                         return ret;
3612
3613                 return perf_sched__timehist(&sched);
3614         } else {
3615                 usage_with_options(sched_usage, sched_options);
3616         }
3617
3618         return 0;
3619 }