execution.cc

   1 #include <stdio.h>
   2 #include <algorithm>
   3 #include <new>
   4 #include <stdarg.h>
   5
   6 #include "model.h"
   7 #include "execution.h"
   8 #include "action.h"
   9 #include "nodestack.h"
  10 #include "schedule.h"
  11 #include "common.h"
  12 #include "clockvector.h"
  13 #include "cyclegraph.h"
  14 #include "datarace.h"
  15 #include "threads-model.h"
  16 #include "bugmessage.h"
  17 #include "fuzzer.h"
  18
  19 #define INITIAL_THREAD_ID       0
  20
  21 /**
  22  * Structure for holding small ModelChecker members that should be snapshotted
  23  */
  24 struct model_snapshot_members {
  25         model_snapshot_members() :
  26                 /* First thread created will have id INITIAL_THREAD_ID */
  27                 next_thread_id(INITIAL_THREAD_ID),
  28                 used_sequence_numbers(0),
  29                 bugs(),
  30                 bad_synchronization(false),
  31                 asserted(false)
  32         { }
  33
  34         ~model_snapshot_members() {
  35                 for (unsigned int i = 0;i < bugs.size();i++)
  36                         delete bugs[i];
  37                 bugs.clear();
  38         }
  39
  40         unsigned int next_thread_id;
  41         modelclock_t used_sequence_numbers;
  42         SnapVector<bug_message *> bugs;
  43         /** @brief Incorrectly-ordered synchronization was made */
  44         bool bad_synchronization;
  45         bool asserted;
  46
  47         SNAPSHOTALLOC
  48 };
  49
  50 /** @brief Constructor */
  51 ModelExecution::ModelExecution(ModelChecker *m, Scheduler *scheduler, NodeStack *node_stack) :
  52         model(m),
  53         params(NULL),
  54         scheduler(scheduler),
  55         action_trace(),
  56         thread_map(2),  /* We'll always need at least 2 threads */
  57         pthread_map(0),
  58         pthread_counter(1),
  59         obj_map(),
  60         condvar_waiters_map(),
  61         obj_thrd_map(),
  62         mutex_map(),
  63         thrd_last_action(1),
  64         thrd_last_fence_release(),
  65         node_stack(node_stack),
  66         priv(new struct model_snapshot_members ()),
  67                          mo_graph(new CycleGraph()),
  68         fuzzer(new Fuzzer())
  69 {
  70         /* Initialize a model-checker thread, for special ModelActions */
  71         model_thread = new Thread(get_next_id());
  72         add_thread(model_thread);
  73         scheduler->register_engine(this);
  74         node_stack->register_engine(this);
  75 }
  76
  77 /** @brief Destructor */
  78 ModelExecution::~ModelExecution()
  79 {
  80         for (unsigned int i = 0;i < get_num_threads();i++)
  81                 delete get_thread(int_to_id(i));
  82
  83         delete mo_graph;
  84         delete priv;
  85 }
  86
  87 int ModelExecution::get_execution_number() const
  88 {
  89         return model->get_execution_number();
  90 }
  91
  92 static action_list_t * get_safe_ptr_action(HashTable<const void *, action_list_t *, uintptr_t, 4> * hash, void * ptr)
  93 {
  94         action_list_t *tmp = hash->get(ptr);
  95         if (tmp == NULL) {
  96                 tmp = new action_list_t();
  97                 hash->put(ptr, tmp);
  98         }
  99         return tmp;
 100 }
 101
 102 static SnapVector<action_list_t> * get_safe_ptr_vect_action(HashTable<void *, SnapVector<action_list_t> *, uintptr_t, 4> * hash, void * ptr)
 103 {
 104         SnapVector<action_list_t> *tmp = hash->get(ptr);
 105         if (tmp == NULL) {
 106                 tmp = new SnapVector<action_list_t>();
 107                 hash->put(ptr, tmp);
 108         }
 109         return tmp;
 110 }
 111
 112 /** @return a thread ID for a new Thread */
 113 thread_id_t ModelExecution::get_next_id()
 114 {
 115         return priv->next_thread_id++;
 116 }
 117
 118 /** @return the number of user threads created during this execution */
 119 unsigned int ModelExecution::get_num_threads() const
 120 {
 121         return priv->next_thread_id;
 122 }
 123
 124 /** @return a sequence number for a new ModelAction */
 125 modelclock_t ModelExecution::get_next_seq_num()
 126 {
 127         return ++priv->used_sequence_numbers;
 128 }
 129
 130 /**
 131  * @brief Should the current action wake up a given thread?
 132  *
 133  * @param curr The current action
 134  * @param thread The thread that we might wake up
 135  * @return True, if we should wake up the sleeping thread; false otherwise
 136  */
 137 bool ModelExecution::should_wake_up(const ModelAction *curr, const Thread *thread) const
 138 {
 139         const ModelAction *asleep = thread->get_pending();
 140         /* Don't allow partial RMW to wake anyone up */
 141         if (curr->is_rmwr())
 142                 return false;
 143         /* Synchronizing actions may have been backtracked */
 144         if (asleep->could_synchronize_with(curr))
 145                 return true;
 146         /* All acquire/release fences and fence-acquire/store-release */
 147         if (asleep->is_fence() && asleep->is_acquire() && curr->is_release())
 148                 return true;
 149         /* Fence-release + store can awake load-acquire on the same location */
 150         if (asleep->is_read() && asleep->is_acquire() && curr->same_var(asleep) && curr->is_write()) {
 151                 ModelAction *fence_release = get_last_fence_release(curr->get_tid());
 152                 if (fence_release && *(get_last_action(thread->get_id())) < *fence_release)
 153                         return true;
 154         }
 155         return false;
 156 }
 157
 158 void ModelExecution::wake_up_sleeping_actions(ModelAction *curr)
 159 {
 160         for (unsigned int i = 0;i < get_num_threads();i++) {
 161                 Thread *thr = get_thread(int_to_id(i));
 162                 if (scheduler->is_sleep_set(thr)) {
 163                         if (should_wake_up(curr, thr))
 164                                 /* Remove this thread from sleep set */
 165                                 scheduler->remove_sleep(thr);
 166                 }
 167         }
 168 }
 169
 170 /** @brief Alert the model-checker that an incorrectly-ordered
 171  * synchronization was made */
 172 void ModelExecution::set_bad_synchronization()
 173 {
 174         priv->bad_synchronization = true;
 175 }
 176
 177 bool ModelExecution::assert_bug(const char *msg)
 178 {
 179         priv->bugs.push_back(new bug_message(msg));
 180
 181         if (isfeasibleprefix()) {
 182                 set_assert();
 183                 return true;
 184         }
 185         return false;
 186 }
 187
 188 /** @return True, if any bugs have been reported for this execution */
 189 bool ModelExecution::have_bug_reports() const
 190 {
 191         return priv->bugs.size() != 0;
 192 }
 193
 194 SnapVector<bug_message *> * ModelExecution::get_bugs() const
 195 {
 196         return &priv->bugs;
 197 }
 198
 199 /**
 200  * Check whether the current trace has triggered an assertion which should halt
 201  * its execution.
 202  *
 203  * @return True, if the execution should be aborted; false otherwise
 204  */
 205 bool ModelExecution::has_asserted() const
 206 {
 207         return priv->asserted;
 208 }
 209
 210 /**
 211  * Trigger a trace assertion which should cause this execution to be halted.
 212  * This can be due to a detected bug or due to an infeasibility that should
 213  * halt ASAP.
 214  */
 215 void ModelExecution::set_assert()
 216 {
 217         priv->asserted = true;
 218 }
 219
 220 /**
 221  * Check if we are in a deadlock. Should only be called at the end of an
 222  * execution, although it should not give false positives in the middle of an
 223  * execution (there should be some ENABLED thread).
 224  *
 225  * @return True if program is in a deadlock; false otherwise
 226  */
 227 bool ModelExecution::is_deadlocked() const
 228 {
 229         bool blocking_threads = false;
 230         for (unsigned int i = 0;i < get_num_threads();i++) {
 231                 thread_id_t tid = int_to_id(i);
 232                 if (is_enabled(tid))
 233                         return false;
 234                 Thread *t = get_thread(tid);
 235                 if (!t->is_model_thread() && t->get_pending())
 236                         blocking_threads = true;
 237         }
 238         return blocking_threads;
 239 }
 240
 241 /**
 242  * Check if this is a complete execution. That is, have all thread completed
 243  * execution (rather than exiting because sleep sets have forced a redundant
 244  * execution).
 245  *
 246  * @return True if the execution is complete.
 247  */
 248 bool ModelExecution::is_complete_execution() const
 249 {
 250         for (unsigned int i = 0;i < get_num_threads();i++)
 251                 if (is_enabled(int_to_id(i)))
 252                         return false;
 253         return true;
 254 }
 255
 256
 257 /**
 258  * Processes a read model action.
 259  * @param curr is the read model action to process.
 260  * @param rf_set is the set of model actions we can possibly read from
 261  * @return True if processing this read updates the mo_graph.
 262  */
 263 void ModelExecution::process_read(ModelAction *curr, SnapVector<ModelAction *> * rf_set)
 264 {
 265         SnapVector<const ModelAction *> * priorset = new SnapVector<const ModelAction *>();
 266         while(true) {
 267
 268                 int index = fuzzer->selectWrite(curr, rf_set);
 269                 ModelAction *rf = (*rf_set)[index];
 270
 271
 272                 ASSERT(rf);
 273                 bool canprune = false;
 274                 if (r_modification_order(curr, rf, priorset, &canprune)) {
 275                         for(unsigned int i=0;i<priorset->size();i++) {
 276                                 mo_graph->addEdge((*priorset)[i], rf);
 277                         }
 278                         read_from(curr, rf);
 279                         get_thread(curr)->set_return_value(curr->get_return_value());
 280                         delete priorset;
 281                         if (canprune && curr->get_type() == ATOMIC_READ) {
 282                                 int tid = id_to_int(curr->get_tid());
 283                                 (*obj_thrd_map.get(curr->get_location()))[tid].pop_back();
 284                         }
 285                         return;
 286                 }
 287                 priorset->clear();
 288                 (*rf_set)[index] = rf_set->back();
 289                 rf_set->pop_back();
 290         }
 291 }
 292
 293 /**
 294  * Processes a lock, trylock, or unlock model action.  @param curr is
 295  * the read model action to process.
 296  *
 297  * The try lock operation checks whether the lock is taken.  If not,
 298  * it falls to the normal lock operation case.  If so, it returns
 299  * fail.
 300  *
 301  * The lock operation has already been checked that it is enabled, so
 302  * it just grabs the lock and synchronizes with the previous unlock.
 303  *
 304  * The unlock operation has to re-enable all of the threads that are
 305  * waiting on the lock.
 306  *
 307  * @return True if synchronization was updated; false otherwise
 308  */
 309 bool ModelExecution::process_mutex(ModelAction *curr)
 310 {
 311         cdsc::mutex *mutex = curr->get_mutex();
 312         struct cdsc::mutex_state *state = NULL;
 313
 314         if (mutex)
 315                 state = mutex->get_state();
 316
 317         switch (curr->get_type()) {
 318         case ATOMIC_TRYLOCK: {
 319                 bool success = !state->locked;
 320                 curr->set_try_lock(success);
 321                 if (!success) {
 322                         get_thread(curr)->set_return_value(0);
 323                         break;
 324                 }
 325                 get_thread(curr)->set_return_value(1);
 326         }
 327         //otherwise fall into the lock case
 328         case ATOMIC_LOCK: {
 329                 if (curr->get_cv()->getClock(state->alloc_tid) <= state->alloc_clock)
 330                         assert_bug("Lock access before initialization");
 331                 state->locked = get_thread(curr);
 332                 ModelAction *unlock = get_last_unlock(curr);
 333                 //synchronize with the previous unlock statement
 334                 if (unlock != NULL) {
 335                         synchronize(unlock, curr);
 336                         return true;
 337                 }
 338                 break;
 339         }
 340         case ATOMIC_WAIT:
 341         case ATOMIC_UNLOCK: {
 342                 //TODO: FIX WAIT SITUATION...WAITS CAN SPURIOUSLY FAIL...TIMED WAITS SHOULD PROBABLY JUST BE THE SAME AS NORMAL WAITS...THINK ABOUT PROBABILITIES THOUGH....AS IN TIMED WAIT MUST FAIL TO GUARANTEE PROGRESS...NORMAL WAIT MAY FAIL...SO NEED NORMAL WAIT TO WORK CORRECTLY IN THE CASE IT SPURIOUSLY FAILS AND IN THE CASE IT DOESN'T...  TIMED WAITS MUST EVENMTUALLY RELEASE...
 343
 344                 /* wake up the other threads */
 345                 for (unsigned int i = 0;i < get_num_threads();i++) {
 346                         Thread *t = get_thread(int_to_id(i));
 347                         Thread *curr_thrd = get_thread(curr);
 348                         if (t->waiting_on() == curr_thrd && t->get_pending()->is_lock())
 349                                 scheduler->wake(t);
 350                 }
 351
 352                 /* unlock the lock - after checking who was waiting on it */
 353                 state->locked = NULL;
 354
 355                 if (!curr->is_wait())
 356                         break;/* The rest is only for ATOMIC_WAIT */
 357
 358                 break;
 359         }
 360         case ATOMIC_NOTIFY_ALL: {
 361                 action_list_t *waiters = get_safe_ptr_action(&condvar_waiters_map, curr->get_location());
 362                 //activate all the waiting threads
 363                 for (action_list_t::iterator rit = waiters->begin();rit != waiters->end();rit++) {
 364                         scheduler->wake(get_thread(*rit));
 365                 }
 366                 waiters->clear();
 367                 break;
 368         }
 369         case ATOMIC_NOTIFY_ONE: {
 370                 action_list_t *waiters = get_safe_ptr_action(&condvar_waiters_map, curr->get_location());
 371                 if (waiters->size() != 0) {
 372                         Thread * thread = fuzzer->selectNotify(waiters);
 373                         scheduler->wake(thread);
 374                 }
 375                 break;
 376         }
 377
 378         default:
 379                 ASSERT(0);
 380         }
 381         return false;
 382 }
 383
 384 /**
 385  * Process a write ModelAction
 386  * @param curr The ModelAction to process
 387  * @return True if the mo_graph was updated or promises were resolved
 388  */
 389 void ModelExecution::process_write(ModelAction *curr)
 390 {
 391
 392         w_modification_order(curr);
 393
 394
 395         get_thread(curr)->set_return_value(VALUE_NONE);
 396 }
 397
 398 /**
 399  * Process a fence ModelAction
 400  * @param curr The ModelAction to process
 401  * @return True if synchronization was updated
 402  */
 403 bool ModelExecution::process_fence(ModelAction *curr)
 404 {
 405         /*
 406          * fence-relaxed: no-op
 407          * fence-release: only log the occurence (not in this function), for
 408          *   use in later synchronization
 409          * fence-acquire (this function): search for hypothetical release
 410          *   sequences
 411          * fence-seq-cst: MO constraints formed in {r,w}_modification_order
 412          */
 413         bool updated = false;
 414         if (curr->is_acquire()) {
 415                 action_list_t *list = &action_trace;
 416                 action_list_t::reverse_iterator rit;
 417                 /* Find X : is_read(X) && X --sb-> curr */
 418                 for (rit = list->rbegin();rit != list->rend();rit++) {
 419                         ModelAction *act = *rit;
 420                         if (act == curr)
 421                                 continue;
 422                         if (act->get_tid() != curr->get_tid())
 423                                 continue;
 424                         /* Stop at the beginning of the thread */
 425                         if (act->is_thread_start())
 426                                 break;
 427                         /* Stop once we reach a prior fence-acquire */
 428                         if (act->is_fence() && act->is_acquire())
 429                                 break;
 430                         if (!act->is_read())
 431                                 continue;
 432                         /* read-acquire will find its own release sequences */
 433                         if (act->is_acquire())
 434                                 continue;
 435
 436                         /* Establish hypothetical release sequences */
 437                         ClockVector *cv = get_hb_from_write(act);
 438                         if (curr->get_cv()->merge(cv))
 439                                 updated = true;
 440                 }
 441         }
 442         return updated;
 443 }
 444
 445 /**
 446  * @brief Process the current action for thread-related activity
 447  *
 448  * Performs current-action processing for a THREAD_* ModelAction. Proccesses
 449  * may include setting Thread status, completing THREAD_FINISH/THREAD_JOIN
 450  * synchronization, etc.  This function is a no-op for non-THREAD actions
 451  * (e.g., ATOMIC_{READ,WRITE,RMW,LOCK}, etc.)
 452  *
 453  * @param curr The current action
 454  * @return True if synchronization was updated or a thread completed
 455  */
 456 bool ModelExecution::process_thread_action(ModelAction *curr)
 457 {
 458         bool updated = false;
 459
 460         switch (curr->get_type()) {
 461         case THREAD_CREATE: {
 462                 thrd_t *thrd = (thrd_t *)curr->get_location();
 463                 struct thread_params *params = (struct thread_params *)curr->get_value();
 464                 Thread *th = new Thread(get_next_id(), thrd, params->func, params->arg, get_thread(curr));
 465                 curr->set_thread_operand(th);
 466                 add_thread(th);
 467                 th->set_creation(curr);
 468                 break;
 469         }
 470         case PTHREAD_CREATE: {
 471                 (*(uint32_t *)curr->get_location()) = pthread_counter++;
 472
 473                 struct pthread_params *params = (struct pthread_params *)curr->get_value();
 474                 Thread *th = new Thread(get_next_id(), NULL, params->func, params->arg, get_thread(curr));
 475                 curr->set_thread_operand(th);
 476                 add_thread(th);
 477                 th->set_creation(curr);
 478
 479                 if ( pthread_map.size() < pthread_counter )
 480                         pthread_map.resize( pthread_counter );
 481                 pthread_map[ pthread_counter-1 ] = th;
 482
 483                 break;
 484         }
 485         case THREAD_JOIN: {
 486                 Thread *blocking = curr->get_thread_operand();
 487                 ModelAction *act = get_last_action(blocking->get_id());
 488                 synchronize(act, curr);
 489                 updated = true; /* trigger rel-seq checks */
 490                 break;
 491         }
 492         case PTHREAD_JOIN: {
 493                 Thread *blocking = curr->get_thread_operand();
 494                 ModelAction *act = get_last_action(blocking->get_id());
 495                 synchronize(act, curr);
 496                 updated = true; /* trigger rel-seq checks */
 497                 break;  // WL: to be add (modified)
 498         }
 499
 500         case THREAD_FINISH: {
 501                 Thread *th = get_thread(curr);
 502                 /* Wake up any joining threads */
 503                 for (unsigned int i = 0;i < get_num_threads();i++) {
 504                         Thread *waiting = get_thread(int_to_id(i));
 505                         if (waiting->waiting_on() == th &&
 506                                         waiting->get_pending()->is_thread_join())
 507                                 scheduler->wake(waiting);
 508                 }
 509                 th->complete();
 510                 updated = true; /* trigger rel-seq checks */
 511                 break;
 512         }
 513         case THREAD_START: {
 514                 break;
 515         }
 516         default:
 517                 break;
 518         }
 519
 520         return updated;
 521 }
 522
 523 /**
 524  * Initialize the current action by performing one or more of the following
 525  * actions, as appropriate: merging RMWR and RMWC/RMW actions, stepping forward
 526  * in the NodeStack, manipulating backtracking sets, allocating and
 527  * initializing clock vectors, and computing the promises to fulfill.
 528  *
 529  * @param curr The current action, as passed from the user context; may be
 530  * freed/invalidated after the execution of this function, with a different
 531  * action "returned" its place (pass-by-reference)
 532  * @return True if curr is a newly-explored action; false otherwise
 533  */
 534 bool ModelExecution::initialize_curr_action(ModelAction **curr)
 535 {
 536         if ((*curr)->is_rmwc() || (*curr)->is_rmw()) {
 537                 ModelAction *newcurr = process_rmw(*curr);
 538                 delete *curr;
 539
 540                 *curr = newcurr;
 541                 return false;
 542         } else {
 543                 ModelAction *newcurr = *curr;
 544
 545                 newcurr->set_seq_number(get_next_seq_num());
 546                 node_stack->add_action(newcurr);
 547                 /* Always compute new clock vector */
 548                 newcurr->create_cv(get_parent_action(newcurr->get_tid()));
 549
 550                 /* Assign most recent release fence */
 551                 newcurr->set_last_fence_release(get_last_fence_release(newcurr->get_tid()));
 552
 553                 return true;    /* This was a new ModelAction */
 554         }
 555 }
 556
 557 /**
 558  * @brief Establish reads-from relation between two actions
 559  *
 560  * Perform basic operations involved with establishing a concrete rf relation,
 561  * including setting the ModelAction data and checking for release sequences.
 562  *
 563  * @param act The action that is reading (must be a read)
 564  * @param rf The action from which we are reading (must be a write)
 565  *
 566  * @return True if this read established synchronization
 567  */
 568
 569 void ModelExecution::read_from(ModelAction *act, ModelAction *rf)
 570 {
 571         ASSERT(rf);
 572         ASSERT(rf->is_write());
 573
 574         act->set_read_from(rf);
 575         if (act->is_acquire()) {
 576                 ClockVector *cv = get_hb_from_write(rf);
 577                 if (cv == NULL)
 578                         return;
 579                 act->get_cv()->merge(cv);
 580         }
 581 }
 582
 583 /**
 584  * @brief Synchronizes two actions
 585  *
 586  * When A synchronizes with B (or A --sw-> B), B inherits A's clock vector.
 587  * This function performs the synchronization as well as providing other hooks
 588  * for other checks along with synchronization.
 589  *
 590  * @param first The left-hand side of the synchronizes-with relation
 591  * @param second The right-hand side of the synchronizes-with relation
 592  * @return True if the synchronization was successful (i.e., was consistent
 593  * with the execution order); false otherwise
 594  */
 595 bool ModelExecution::synchronize(const ModelAction *first, ModelAction *second)
 596 {
 597         if (*second < *first) {
 598                 set_bad_synchronization();
 599                 return false;
 600         }
 601         return second->synchronize_with(first);
 602 }
 603
 604 /**
 605  * @brief Check whether a model action is enabled.
 606  *
 607  * Checks whether an operation would be successful (i.e., is a lock already
 608  * locked, or is the joined thread already complete).
 609  *
 610  * For yield-blocking, yields are never enabled.
 611  *
 612  * @param curr is the ModelAction to check whether it is enabled.
 613  * @return a bool that indicates whether the action is enabled.
 614  */
 615 bool ModelExecution::check_action_enabled(ModelAction *curr) {
 616         if (curr->is_lock()) {
 617                 cdsc::mutex *lock = curr->get_mutex();
 618                 struct cdsc::mutex_state *state = lock->get_state();
 619                 if (state->locked)
 620                         return false;
 621         } else if (curr->is_thread_join()) {
 622                 Thread *blocking = curr->get_thread_operand();
 623                 if (!blocking->is_complete()) {
 624                         return false;
 625                 }
 626         }
 627
 628         return true;
 629 }
 630
 631 /**
 632  * This is the heart of the model checker routine. It performs model-checking
 633  * actions corresponding to a given "current action." Among other processes, it
 634  * calculates reads-from relationships, updates synchronization clock vectors,
 635  * forms a memory_order constraints graph, and handles replay/backtrack
 636  * execution when running permutations of previously-observed executions.
 637  *
 638  * @param curr The current action to process
 639  * @return The ModelAction that is actually executed; may be different than
 640  * curr
 641  */
 642 ModelAction * ModelExecution::check_current_action(ModelAction *curr)
 643 {
 644         ASSERT(curr);
 645         bool second_part_of_rmw = curr->is_rmwc() || curr->is_rmw();
 646         bool newly_explored = initialize_curr_action(&curr);
 647
 648         DBG();
 649
 650         wake_up_sleeping_actions(curr);
 651
 652         /* Add the action to lists before any other model-checking tasks */
 653         if (!second_part_of_rmw && curr->get_type() != NOOP)
 654                 add_action_to_lists(curr);
 655
 656         SnapVector<ModelAction *> * rf_set = NULL;
 657         /* Build may_read_from set for newly-created actions */
 658         if (newly_explored && curr->is_read())
 659                 rf_set = build_may_read_from(curr);
 660
 661         process_thread_action(curr);
 662
 663         if (curr->is_read() && !second_part_of_rmw) {
 664                 process_read(curr, rf_set);
 665                 delete rf_set;
 666         } else {
 667                 ASSERT(rf_set == NULL);
 668         }
 669
 670         if (curr->is_write())
 671                 process_write(curr);
 672
 673         if (curr->is_fence())
 674                 process_fence(curr);
 675
 676         if (curr->is_mutex_op())
 677                 process_mutex(curr);
 678
 679         return curr;
 680 }
 681
 682 /**
 683  * This is the strongest feasibility check available.
 684  * @return whether the current trace (partial or complete) must be a prefix of
 685  * a feasible trace.
 686  */
 687 bool ModelExecution::isfeasibleprefix() const
 688 {
 689         return !is_infeasible();
 690 }
 691
 692 /**
 693  * Print disagnostic information about an infeasible execution
 694  * @param prefix A string to prefix the output with; if NULL, then a default
 695  * message prefix will be provided
 696  */
 697 void ModelExecution::print_infeasibility(const char *prefix) const
 698 {
 699         char buf[100];
 700         char *ptr = buf;
 701         if (priv->bad_synchronization)
 702                 ptr += sprintf(ptr, "[bad sw ordering]");
 703         if (ptr != buf)
 704                 model_print("%s: %s", prefix ? prefix : "Infeasible", buf);
 705 }
 706
 707 /**
 708  * Check if the current partial trace is infeasible. Does not check any
 709  * end-of-execution flags, which might rule out the execution. Thus, this is
 710  * useful only for ruling an execution as infeasible.
 711  * @return whether the current partial trace is infeasible.
 712  */
 713 bool ModelExecution::is_infeasible() const
 714 {
 715         return priv->bad_synchronization;
 716 }
 717
 718 /** Close out a RMWR by converting previous RMWR into a RMW or READ. */
 719 ModelAction * ModelExecution::process_rmw(ModelAction *act) {
 720         ModelAction *lastread = get_last_action(act->get_tid());
 721         lastread->process_rmw(act);
 722         if (act->is_rmw()) {
 723                 mo_graph->addRMWEdge(lastread->get_reads_from(), lastread);
 724         }
 725         return lastread;
 726 }
 727
 728 /**
 729  * @brief Updates the mo_graph with the constraints imposed from the current
 730  * read.
 731  *
 732  * Basic idea is the following: Go through each other thread and find
 733  * the last action that happened before our read.  Two cases:
 734  *
 735  * -# The action is a write: that write must either occur before
 736  * the write we read from or be the write we read from.
 737  * -# The action is a read: the write that that action read from
 738  * must occur before the write we read from or be the same write.
 739  *
 740  * @param curr The current action. Must be a read.
 741  * @param rf The ModelAction or Promise that curr reads from. Must be a write.
 742  * @return True if modification order edges were added; false otherwise
 743  */
 744
 745 bool ModelExecution::r_modification_order(ModelAction *curr, const ModelAction *rf, SnapVector<const ModelAction *> * priorset, bool * canprune)
 746 {
 747         SnapVector<action_list_t> *thrd_lists = obj_thrd_map.get(curr->get_location());
 748         unsigned int i;
 749         ASSERT(curr->is_read());
 750
 751         /* Last SC fence in the current thread */
 752         ModelAction *last_sc_fence_local = get_last_seq_cst_fence(curr->get_tid(), NULL);
 753
 754         int tid = curr->get_tid();
 755         ModelAction *prev_same_thread = NULL;
 756         /* Iterate over all threads */
 757         for (i = 0;i < thrd_lists->size();i++, tid = (((unsigned int)(tid+1)) == thrd_lists->size()) ? 0 : tid + 1) {
 758                 /* Last SC fence in thread tid */
 759                 ModelAction *last_sc_fence_thread_local = NULL;
 760                 if (i != 0)
 761                         last_sc_fence_thread_local = get_last_seq_cst_fence(int_to_id(tid), NULL);
 762
 763                 /* Last SC fence in thread tid, before last SC fence in current thread */
 764                 ModelAction *last_sc_fence_thread_before = NULL;
 765                 if (last_sc_fence_local)
 766                         last_sc_fence_thread_before = get_last_seq_cst_fence(int_to_id(tid), last_sc_fence_local);
 767
 768                 //Only need to iterate if either hb has changed for thread in question or SC fence after last operation...
 769                 if (prev_same_thread != NULL &&
 770                                 (prev_same_thread->get_cv()->getClock(tid) == curr->get_cv()->getClock(tid)) &&
 771                                 (last_sc_fence_thread_local == NULL || *last_sc_fence_thread_local < *prev_same_thread)) {
 772                         continue;
 773                 }
 774
 775                 /* Iterate over actions in thread, starting from most recent */
 776                 action_list_t *list = &(*thrd_lists)[tid];
 777                 action_list_t::reverse_iterator rit;
 778                 for (rit = list->rbegin();rit != list->rend();rit++) {
 779                         ModelAction *act = *rit;
 780
 781                         /* Skip curr */
 782                         if (act == curr)
 783                                 continue;
 784                         /* Don't want to add reflexive edges on 'rf' */
 785                         if (act->equals(rf)) {
 786                                 if (act->happens_before(curr))
 787                                         break;
 788                                 else
 789                                         continue;
 790                         }
 791
 792                         if (act->is_write()) {
 793                                 /* C++, Section 29.3 statement 5 */
 794                                 if (curr->is_seqcst() && last_sc_fence_thread_local &&
 795                                                 *act < *last_sc_fence_thread_local) {
 796                                         if (mo_graph->checkReachable(rf, act))
 797                                                 return false;
 798                                         priorset->push_back(act);
 799                                         break;
 800                                 }
 801                                 /* C++, Section 29.3 statement 4 */
 802                                 else if (act->is_seqcst() && last_sc_fence_local &&
 803                                                                  *act < *last_sc_fence_local) {
 804                                         if (mo_graph->checkReachable(rf, act))
 805                                                 return false;
 806                                         priorset->push_back(act);
 807                                         break;
 808                                 }
 809                                 /* C++, Section 29.3 statement 6 */
 810                                 else if (last_sc_fence_thread_before &&
 811                                                                  *act < *last_sc_fence_thread_before) {
 812                                         if (mo_graph->checkReachable(rf, act))
 813                                                 return false;
 814                                         priorset->push_back(act);
 815                                         break;
 816                                 }
 817                         }
 818
 819                         /*
 820                          * Include at most one act per-thread that "happens
 821                          * before" curr
 822                          */
 823                         if (act->happens_before(curr)) {
 824                                 if (i==0) {
 825                                         if (last_sc_fence_local == NULL ||
 826                                                         (*last_sc_fence_local < *prev_same_thread)) {
 827                                                 prev_same_thread = act;
 828                                         }
 829                                 }
 830                                 if (act->is_write()) {
 831                                         if (mo_graph->checkReachable(rf, act))
 832                                                 return false;
 833                                         priorset->push_back(act);
 834                                 } else {
 835                                         const ModelAction *prevrf = act->get_reads_from();
 836                                         if (!prevrf->equals(rf)) {
 837                                                 if (mo_graph->checkReachable(rf, prevrf))
 838                                                         return false;
 839                                                 priorset->push_back(prevrf);
 840                                         } else {
 841                                                 if (act->get_tid() == curr->get_tid()) {
 842                                                         //Can prune curr from obj list
 843                                                         *canprune = true;
 844                                                 }
 845                                         }
 846                                 }
 847                                 break;
 848                         }
 849                 }
 850         }
 851         return true;
 852 }
 853
 854 /**
 855  * Updates the mo_graph with the constraints imposed from the current write.
 856  *
 857  * Basic idea is the following: Go through each other thread and find
 858  * the lastest action that happened before our write.  Two cases:
 859  *
 860  * (1) The action is a write => that write must occur before
 861  * the current write
 862  *
 863  * (2) The action is a read => the write that that action read from
 864  * must occur before the current write.
 865  *
 866  * This method also handles two other issues:
 867  *
 868  * (I) Sequential Consistency: Making sure that if the current write is
 869  * seq_cst, that it occurs after the previous seq_cst write.
 870  *
 871  * (II) Sending the write back to non-synchronizing reads.
 872  *
 873  * @param curr The current action. Must be a write.
 874  * @param send_fv A vector for stashing reads to which we may pass our future
 875  * value. If NULL, then don't record any future values.
 876  * @return True if modification order edges were added; false otherwise
 877  */
 878 void ModelExecution::w_modification_order(ModelAction *curr)
 879 {
 880         SnapVector<action_list_t> *thrd_lists = obj_thrd_map.get(curr->get_location());
 881         unsigned int i;
 882         ASSERT(curr->is_write());
 883
 884         if (curr->is_seqcst()) {
 885                 /* We have to at least see the last sequentially consistent write,
 886                          so we are initialized. */
 887                 ModelAction *last_seq_cst = get_last_seq_cst_write(curr);
 888                 if (last_seq_cst != NULL) {
 889                         mo_graph->addEdge(last_seq_cst, curr);
 890                 }
 891         }
 892
 893         /* Last SC fence in the current thread */
 894         ModelAction *last_sc_fence_local = get_last_seq_cst_fence(curr->get_tid(), NULL);
 895
 896         /* Iterate over all threads */
 897         for (i = 0;i < thrd_lists->size();i++) {
 898                 /* Last SC fence in thread i, before last SC fence in current thread */
 899                 ModelAction *last_sc_fence_thread_before = NULL;
 900                 if (last_sc_fence_local && int_to_id((int)i) != curr->get_tid())
 901                         last_sc_fence_thread_before = get_last_seq_cst_fence(int_to_id(i), last_sc_fence_local);
 902
 903                 /* Iterate over actions in thread, starting from most recent */
 904                 action_list_t *list = &(*thrd_lists)[i];
 905                 action_list_t::reverse_iterator rit;
 906                 bool force_edge = false;
 907                 for (rit = list->rbegin();rit != list->rend();rit++) {
 908                         ModelAction *act = *rit;
 909                         if (act == curr) {
 910                                 /*
 911                                  * 1) If RMW and it actually read from something, then we
 912                                  * already have all relevant edges, so just skip to next
 913                                  * thread.
 914                                  *
 915                                  * 2) If RMW and it didn't read from anything, we should
 916                                  * whatever edge we can get to speed up convergence.
 917                                  *
 918                                  * 3) If normal write, we need to look at earlier actions, so
 919                                  * continue processing list.
 920                                  */
 921                                 force_edge = true;
 922                                 if (curr->is_rmw()) {
 923                                         if (curr->get_reads_from() != NULL)
 924                                                 break;
 925                                         else
 926                                                 continue;
 927                                 } else
 928                                         continue;
 929                         }
 930
 931                         /* C++, Section 29.3 statement 7 */
 932                         if (last_sc_fence_thread_before && act->is_write() &&
 933                                         *act < *last_sc_fence_thread_before) {
 934                                 mo_graph->addEdge(act, curr, force_edge);
 935                                 break;
 936                         }
 937
 938                         /*
 939                          * Include at most one act per-thread that "happens
 940                          * before" curr
 941                          */
 942                         if (act->happens_before(curr)) {
 943                                 /*
 944                                  * Note: if act is RMW, just add edge:
 945                                  *   act --mo--> curr
 946                                  * The following edge should be handled elsewhere:
 947                                  *   readfrom(act) --mo--> act
 948                                  */
 949                                 if (act->is_write())
 950                                         mo_graph->addEdge(act, curr, force_edge);
 951                                 else if (act->is_read()) {
 952                                         //if previous read accessed a null, just keep going
 953                                         mo_graph->addEdge(act->get_reads_from(), curr, force_edge);
 954                                 }
 955                                 break;
 956                         } else if (act->is_read() && !act->could_synchronize_with(curr) &&
 957                                                                  !act->same_thread(curr)) {
 958                                 /* We have an action that:
 959                                    (1) did not happen before us
 960                                    (2) is a read and we are a write
 961                                    (3) cannot synchronize with us
 962                                    (4) is in a different thread
 963                                    =>
 964                                    that read could potentially read from our write.  Note that
 965                                    these checks are overly conservative at this point, we'll
 966                                    do more checks before actually removing the
 967                                    pendingfuturevalue.
 968
 969                                  */
 970
 971                         }
 972                 }
 973         }
 974 }
 975
 976 /**
 977  * Arbitrary reads from the future are not allowed. Section 29.3 part 9 places
 978  * some constraints. This method checks one the following constraint (others
 979  * require compiler support):
 980  *
 981  *   If X --hb-> Y --mo-> Z, then X should not read from Z.
 982  *   If X --hb-> Y, A --rf-> Y, and A --mo-> Z, then X should not read from Z.
 983  */
 984 bool ModelExecution::mo_may_allow(const ModelAction *writer, const ModelAction *reader)
 985 {
 986         SnapVector<action_list_t> *thrd_lists = obj_thrd_map.get(reader->get_location());
 987         unsigned int i;
 988         /* Iterate over all threads */
 989         for (i = 0;i < thrd_lists->size();i++) {
 990                 const ModelAction *write_after_read = NULL;
 991
 992                 /* Iterate over actions in thread, starting from most recent */
 993                 action_list_t *list = &(*thrd_lists)[i];
 994                 action_list_t::reverse_iterator rit;
 995                 for (rit = list->rbegin();rit != list->rend();rit++) {
 996                         ModelAction *act = *rit;
 997
 998                         /* Don't disallow due to act == reader */
 999                         if (!reader->happens_before(act) || reader == act)
1000                                 break;
1001                         else if (act->is_write())
1002                                 write_after_read = act;
1003                         else if (act->is_read() && act->get_reads_from() != NULL)
1004                                 write_after_read = act->get_reads_from();
1005                 }
1006
1007                 if (write_after_read && write_after_read != writer && mo_graph->checkReachable(write_after_read, writer))
1008                         return false;
1009         }
1010         return true;
1011 }
1012
1013 /**
1014  * Computes the clock vector that happens before propagates from this write.
1015  *
1016  * @param rf The action that might be part of a release sequence. Must be a
1017  * write.
1018  * @return ClockVector of happens before relation.
1019  */
1020
1021 ClockVector * ModelExecution::get_hb_from_write(ModelAction *rf) const {
1022         SnapVector<const ModelAction *> * processset = NULL;
1023         for ( ;rf != NULL;rf = rf->get_reads_from()) {
1024                 ASSERT(rf->is_write());
1025                 if (!rf->is_rmw() || (rf->is_acquire() && rf->is_release()) || rf->get_rfcv() != NULL)
1026                         break;
1027                 if (processset == NULL)
1028                         processset = new SnapVector<const ModelAction *>();
1029                 processset->push_back(rf);
1030         }
1031
1032         int i = (processset == NULL) ? 1 : processset->size();
1033
1034         ClockVector * vec = NULL;
1035         for(;i > 0 ;i--) {
1036                 if (rf->get_rfcv() != NULL) {
1037                         vec = rf->get_rfcv();
1038                 } else if (rf->is_acquire() && rf->is_release()) {
1039                         vec = rf->get_cv();
1040                 } else if (rf->is_release() && !rf->is_rmw()) {
1041                         vec = rf->get_cv();
1042                 } else if (rf->is_release()) {
1043                         //have rmw that is release and doesn't have a rfcv
1044                         (vec = new ClockVector(vec, NULL))->merge(rf->get_cv());
1045                         rf->set_rfcv(vec);
1046                 } else {
1047                         //operation that isn't release
1048                         if (rf->get_last_fence_release()) {
1049                                 if (vec == NULL)
1050                                         vec = rf->get_last_fence_release()->get_cv();
1051                                 else
1052                                         (vec=new ClockVector(vec, NULL))->merge(rf->get_last_fence_release()->get_cv());
1053                         }
1054                         rf->set_rfcv(vec);
1055                 }
1056         }
1057         if (processset != NULL)
1058                 delete processset;
1059         return vec;
1060 }
1061
1062 /**
1063  * Performs various bookkeeping operations for the current ModelAction. For
1064  * instance, adds action to the per-object, per-thread action vector and to the
1065  * action trace list of all thread actions.
1066  *
1067  * @param act is the ModelAction to add.
1068  */
1069 void ModelExecution::add_action_to_lists(ModelAction *act)
1070 {
1071         int tid = id_to_int(act->get_tid());
1072         ModelAction *uninit = NULL;
1073         int uninit_id = -1;
1074         action_list_t *list = get_safe_ptr_action(&obj_map, act->get_location());
1075         if (list->empty() && act->is_atomic_var()) {
1076                 uninit = get_uninitialized_action(act);
1077                 uninit_id = id_to_int(uninit->get_tid());
1078                 list->push_front(uninit);
1079         }
1080         list->push_back(act);
1081
1082         action_trace.push_back(act);
1083         if (uninit)
1084                 action_trace.push_front(uninit);
1085
1086         SnapVector<action_list_t> *vec = get_safe_ptr_vect_action(&obj_thrd_map, act->get_location());
1087         if (tid >= (int)vec->size())
1088                 vec->resize(priv->next_thread_id);
1089         (*vec)[tid].push_back(act);
1090         if (uninit)
1091                 (*vec)[uninit_id].push_front(uninit);
1092
1093         if ((int)thrd_last_action.size() <= tid)
1094                 thrd_last_action.resize(get_num_threads());
1095         thrd_last_action[tid] = act;
1096         if (uninit)
1097                 thrd_last_action[uninit_id] = uninit;
1098
1099         if (act->is_fence() && act->is_release()) {
1100                 if ((int)thrd_last_fence_release.size() <= tid)
1101                         thrd_last_fence_release.resize(get_num_threads());
1102                 thrd_last_fence_release[tid] = act;
1103         }
1104
1105         if (act->is_wait()) {
1106                 void *mutex_loc = (void *) act->get_value();
1107                 get_safe_ptr_action(&obj_map, mutex_loc)->push_back(act);
1108
1109                 SnapVector<action_list_t> *vec = get_safe_ptr_vect_action(&obj_thrd_map, mutex_loc);
1110                 if (tid >= (int)vec->size())
1111                         vec->resize(priv->next_thread_id);
1112                 (*vec)[tid].push_back(act);
1113         }
1114 }
1115
1116 /**
1117  * @brief Get the last action performed by a particular Thread
1118  * @param tid The thread ID of the Thread in question
1119  * @return The last action in the thread
1120  */
1121 ModelAction * ModelExecution::get_last_action(thread_id_t tid) const
1122 {
1123         int threadid = id_to_int(tid);
1124         if (threadid < (int)thrd_last_action.size())
1125                 return thrd_last_action[id_to_int(tid)];
1126         else
1127                 return NULL;
1128 }
1129
1130 /**
1131  * @brief Get the last fence release performed by a particular Thread
1132  * @param tid The thread ID of the Thread in question
1133  * @return The last fence release in the thread, if one exists; NULL otherwise
1134  */
1135 ModelAction * ModelExecution::get_last_fence_release(thread_id_t tid) const
1136 {
1137         int threadid = id_to_int(tid);
1138         if (threadid < (int)thrd_last_fence_release.size())
1139                 return thrd_last_fence_release[id_to_int(tid)];
1140         else
1141                 return NULL;
1142 }
1143
1144 /**
1145  * Gets the last memory_order_seq_cst write (in the total global sequence)
1146  * performed on a particular object (i.e., memory location), not including the
1147  * current action.
1148  * @param curr The current ModelAction; also denotes the object location to
1149  * check
1150  * @return The last seq_cst write
1151  */
1152 ModelAction * ModelExecution::get_last_seq_cst_write(ModelAction *curr) const
1153 {
1154         void *location = curr->get_location();
1155         action_list_t *list = obj_map.get(location);
1156         /* Find: max({i in dom(S) | seq_cst(t_i) && isWrite(t_i) && samevar(t_i, t)}) */
1157         action_list_t::reverse_iterator rit;
1158         for (rit = list->rbegin();(*rit) != curr;rit++)
1159                 ;
1160         rit++;  /* Skip past curr */
1161         for ( ;rit != list->rend();rit++)
1162                 if ((*rit)->is_write() && (*rit)->is_seqcst())
1163                         return *rit;
1164         return NULL;
1165 }
1166
1167 /**
1168  * Gets the last memory_order_seq_cst fence (in the total global sequence)
1169  * performed in a particular thread, prior to a particular fence.
1170  * @param tid The ID of the thread to check
1171  * @param before_fence The fence from which to begin the search; if NULL, then
1172  * search for the most recent fence in the thread.
1173  * @return The last prior seq_cst fence in the thread, if exists; otherwise, NULL
1174  */
1175 ModelAction * ModelExecution::get_last_seq_cst_fence(thread_id_t tid, const ModelAction *before_fence) const
1176 {
1177         /* All fences should have location FENCE_LOCATION */
1178         action_list_t *list = obj_map.get(FENCE_LOCATION);
1179
1180         if (!list)
1181                 return NULL;
1182
1183         action_list_t::reverse_iterator rit = list->rbegin();
1184
1185         if (before_fence) {
1186                 for (;rit != list->rend();rit++)
1187                         if (*rit == before_fence)
1188                                 break;
1189
1190                 ASSERT(*rit == before_fence);
1191                 rit++;
1192         }
1193
1194         for (;rit != list->rend();rit++)
1195                 if ((*rit)->is_fence() && (tid == (*rit)->get_tid()) && (*rit)->is_seqcst())
1196                         return *rit;
1197         return NULL;
1198 }
1199
1200 /**
1201  * Gets the last unlock operation performed on a particular mutex (i.e., memory
1202  * location). This function identifies the mutex according to the current
1203  * action, which is presumed to perform on the same mutex.
1204  * @param curr The current ModelAction; also denotes the object location to
1205  * check
1206  * @return The last unlock operation
1207  */
1208 ModelAction * ModelExecution::get_last_unlock(ModelAction *curr) const
1209 {
1210         void *location = curr->get_location();
1211
1212         action_list_t *list = obj_map.get(location);
1213         /* Find: max({i in dom(S) | isUnlock(t_i) && samevar(t_i, t)}) */
1214         action_list_t::reverse_iterator rit;
1215         for (rit = list->rbegin();rit != list->rend();rit++)
1216                 if ((*rit)->is_unlock() || (*rit)->is_wait())
1217                         return *rit;
1218         return NULL;
1219 }
1220
1221 ModelAction * ModelExecution::get_parent_action(thread_id_t tid) const
1222 {
1223         ModelAction *parent = get_last_action(tid);
1224         if (!parent)
1225                 parent = get_thread(tid)->get_creation();
1226         return parent;
1227 }
1228
1229 /**
1230  * Returns the clock vector for a given thread.
1231  * @param tid The thread whose clock vector we want
1232  * @return Desired clock vector
1233  */
1234 ClockVector * ModelExecution::get_cv(thread_id_t tid) const
1235 {
1236         ModelAction *firstaction=get_parent_action(tid);
1237         return firstaction != NULL ? firstaction->get_cv() : NULL;
1238 }
1239
1240 bool valequals(uint64_t val1, uint64_t val2, int size) {
1241         switch(size) {
1242         case 1:
1243                 return ((uint8_t)val1) == ((uint8_t)val2);
1244         case 2:
1245                 return ((uint16_t)val1) == ((uint16_t)val2);
1246         case 4:
1247                 return ((uint32_t)val1) == ((uint32_t)val2);
1248         case 8:
1249                 return val1==val2;
1250         default:
1251                 ASSERT(0);
1252                 return false;
1253         }
1254 }
1255
1256 /**
1257  * Build up an initial set of all past writes that this 'read' action may read
1258  * from, as well as any previously-observed future values that must still be valid.
1259  *
1260  * @param curr is the current ModelAction that we are exploring; it must be a
1261  * 'read' operation.
1262  */
1263 SnapVector<ModelAction *> *  ModelExecution::build_may_read_from(ModelAction *curr)
1264 {
1265         SnapVector<action_list_t> *thrd_lists = obj_thrd_map.get(curr->get_location());
1266         unsigned int i;
1267         ASSERT(curr->is_read());
1268
1269         ModelAction *last_sc_write = NULL;
1270
1271         if (curr->is_seqcst())
1272                 last_sc_write = get_last_seq_cst_write(curr);
1273
1274         SnapVector<ModelAction *> * rf_set = new SnapVector<ModelAction *>();
1275
1276         /* Iterate over all threads */
1277         for (i = 0;i < thrd_lists->size();i++) {
1278                 /* Iterate over actions in thread, starting from most recent */
1279                 action_list_t *list = &(*thrd_lists)[i];
1280                 action_list_t::reverse_iterator rit;
1281                 for (rit = list->rbegin();rit != list->rend();rit++) {
1282                         ModelAction *act = *rit;
1283
1284                         /* Only consider 'write' actions */
1285                         if (!act->is_write()) {
1286                                 if (act != curr && act->is_read() && act->happens_before(curr)) {
1287                                         ModelAction *tmp = act->get_reads_from();
1288                                         if (((unsigned int) id_to_int(tmp->get_tid()))==i)
1289                                                 act = tmp;
1290                                         else
1291                                                 break;
1292                                 } else
1293                                         continue;
1294                         }
1295
1296                         if (act == curr)
1297                                 continue;
1298
1299                         /* Don't consider more than one seq_cst write if we are a seq_cst read. */
1300                         bool allow_read = true;
1301
1302                         if (curr->is_seqcst() && (act->is_seqcst() || (last_sc_write != NULL && act->happens_before(last_sc_write))) && act != last_sc_write)
1303                                 allow_read = false;
1304
1305                         /* Need to check whether we will have two RMW reading from the same value */
1306                         if (curr->is_rmwr()) {
1307                                 /* It is okay if we have a failing CAS */
1308                                 if (!curr->is_rmwrcas() ||
1309                                                 valequals(curr->get_value(), act->get_value(), curr->getSize())) {
1310                                         //Need to make sure we aren't the second RMW
1311                                         CycleNode * node = mo_graph->getNode_noCreate(act);
1312                                         if (node != NULL && node->getRMW() != NULL) {
1313                                                 //we are the second RMW
1314                                                 allow_read = false;
1315                                         }
1316                                 }
1317                         }
1318
1319                         if (allow_read) {
1320                                 /* Only add feasible reads */
1321                                 rf_set->push_back(act);
1322                         }
1323
1324                         /* Include at most one act per-thread that "happens before" curr */
1325                         if (act->happens_before(curr))
1326                                 break;
1327                 }
1328         }
1329
1330         if (DBG_ENABLED()) {
1331                 model_print("Reached read action:\n");
1332                 curr->print();
1333                 model_print("End printing read_from_past\n");
1334         }
1335         return rf_set;
1336 }
1337
1338 /**
1339  * @brief Get an action representing an uninitialized atomic
1340  *
1341  * This function may create a new one or try to retrieve one from the NodeStack
1342  *
1343  * @param curr The current action, which prompts the creation of an UNINIT action
1344  * @return A pointer to the UNINIT ModelAction
1345  */
1346 ModelAction * ModelExecution::get_uninitialized_action(const ModelAction *curr) const
1347 {
1348         Node *node = curr->get_node();
1349         ModelAction *act = node->get_uninit_action();
1350         if (!act) {
1351                 act = new ModelAction(ATOMIC_UNINIT, std::memory_order_relaxed, curr->get_location(), params->uninitvalue, model_thread);
1352                 node->set_uninit_action(act);
1353         }
1354         act->create_cv(NULL);
1355         return act;
1356 }
1357
1358 static void print_list(const action_list_t *list)
1359 {
1360         action_list_t::const_iterator it;
1361
1362         model_print("------------------------------------------------------------------------------------\n");
1363         model_print("#    t    Action type     MO       Location         Value               Rf  CV\n");
1364         model_print("------------------------------------------------------------------------------------\n");
1365
1366         unsigned int hash = 0;
1367
1368         for (it = list->begin();it != list->end();it++) {
1369                 const ModelAction *act = *it;
1370                 if (act->get_seq_number() > 0)
1371                         act->print();
1372                 hash = hash^(hash<<3)^((*it)->hash());
1373         }
1374         model_print("HASH %u\n", hash);
1375         model_print("------------------------------------------------------------------------------------\n");
1376 }
1377
1378 #if SUPPORT_MOD_ORDER_DUMP
1379 void ModelExecution::dumpGraph(char *filename) const
1380 {
1381         char buffer[200];
1382         sprintf(buffer, "%s.dot", filename);
1383         FILE *file = fopen(buffer, "w");
1384         fprintf(file, "digraph %s {\n", filename);
1385         mo_graph->dumpNodes(file);
1386         ModelAction **thread_array = (ModelAction **)model_calloc(1, sizeof(ModelAction *) * get_num_threads());
1387
1388         for (action_list_t::const_iterator it = action_trace.begin();it != action_trace.end();it++) {
1389                 ModelAction *act = *it;
1390                 if (act->is_read()) {
1391                         mo_graph->dot_print_node(file, act);
1392                         mo_graph->dot_print_edge(file,
1393                                                                                                                          act->get_reads_from(),
1394                                                                                                                          act,
1395                                                                                                                          "label=\"rf\", color=red, weight=2");
1396                 }
1397                 if (thread_array[act->get_tid()]) {
1398                         mo_graph->dot_print_edge(file,
1399                                                                                                                          thread_array[id_to_int(act->get_tid())],
1400                                                                                                                          act,
1401                                                                                                                          "label=\"sb\", color=blue, weight=400");
1402                 }
1403
1404                 thread_array[act->get_tid()] = act;
1405         }
1406         fprintf(file, "}\n");
1407         model_free(thread_array);
1408         fclose(file);
1409 }
1410 #endif
1411
1412 /** @brief Prints an execution trace summary. */
1413 void ModelExecution::print_summary() const
1414 {
1415 #if SUPPORT_MOD_ORDER_DUMP
1416         char buffername[100];
1417         sprintf(buffername, "exec%04u", get_execution_number());
1418         mo_graph->dumpGraphToFile(buffername);
1419         sprintf(buffername, "graph%04u", get_execution_number());
1420         dumpGraph(buffername);
1421 #endif
1422
1423         model_print("Execution trace %d:", get_execution_number());
1424         if (isfeasibleprefix()) {
1425                 if (scheduler->all_threads_sleeping())
1426                         model_print(" SLEEP-SET REDUNDANT");
1427                 if (have_bug_reports())
1428                         model_print(" DETECTED BUG(S)");
1429         } else
1430                 print_infeasibility(" INFEASIBLE");
1431         model_print("\n");
1432
1433         print_list(&action_trace);
1434         model_print("\n");
1435
1436 }
1437
1438 /**
1439  * Add a Thread to the system for the first time. Should only be called once
1440  * per thread.
1441  * @param t The Thread to add
1442  */
1443 void ModelExecution::add_thread(Thread *t)
1444 {
1445         unsigned int i = id_to_int(t->get_id());
1446         if (i >= thread_map.size())
1447                 thread_map.resize(i + 1);
1448         thread_map[i] = t;
1449         if (!t->is_model_thread())
1450                 scheduler->add_thread(t);
1451 }
1452
1453 /**
1454  * @brief Get a Thread reference by its ID
1455  * @param tid The Thread's ID
1456  * @return A Thread reference
1457  */
1458 Thread * ModelExecution::get_thread(thread_id_t tid) const
1459 {
1460         unsigned int i = id_to_int(tid);
1461         if (i < thread_map.size())
1462                 return thread_map[i];
1463         return NULL;
1464 }
1465
1466 /**
1467  * @brief Get a reference to the Thread in which a ModelAction was executed
1468  * @param act The ModelAction
1469  * @return A Thread reference
1470  */
1471 Thread * ModelExecution::get_thread(const ModelAction *act) const
1472 {
1473         return get_thread(act->get_tid());
1474 }
1475
1476 /**
1477  * @brief Get a Thread reference by its pthread ID
1478  * @param index The pthread's ID
1479  * @return A Thread reference
1480  */
1481 Thread * ModelExecution::get_pthread(pthread_t pid) {
1482         union {
1483                 pthread_t p;
1484                 uint32_t v;
1485         } x;
1486         x.p = pid;
1487         uint32_t thread_id = x.v;
1488         if (thread_id < pthread_counter + 1) return pthread_map[thread_id];
1489         else return NULL;
1490 }
1491
1492 /**
1493  * @brief Check if a Thread is currently enabled
1494  * @param t The Thread to check
1495  * @return True if the Thread is currently enabled
1496  */
1497 bool ModelExecution::is_enabled(Thread *t) const
1498 {
1499         return scheduler->is_enabled(t);
1500 }
1501
1502 /**
1503  * @brief Check if a Thread is currently enabled
1504  * @param tid The ID of the Thread to check
1505  * @return True if the Thread is currently enabled
1506  */
1507 bool ModelExecution::is_enabled(thread_id_t tid) const
1508 {
1509         return scheduler->is_enabled(tid);
1510 }
1511
1512 /**
1513  * @brief Select the next thread to execute based on the curren action
1514  *
1515  * RMW actions occur in two parts, and we cannot split them. And THREAD_CREATE
1516  * actions should be followed by the execution of their child thread. In either
1517  * case, the current action should determine the next thread schedule.
1518  *
1519  * @param curr The current action
1520  * @return The next thread to run, if the current action will determine this
1521  * selection; otherwise NULL
1522  */
1523 Thread * ModelExecution::action_select_next_thread(const ModelAction *curr) const
1524 {
1525         /* Do not split atomic RMW */
1526         if (curr->is_rmwr())
1527                 return get_thread(curr);
1528         /* Follow CREATE with the created thread */
1529         /* which is not needed, because model.cc takes care of this */
1530         if (curr->get_type() == THREAD_CREATE)
1531                 return curr->get_thread_operand();
1532         if (curr->get_type() == PTHREAD_CREATE) {
1533                 return curr->get_thread_operand();
1534         }
1535         return NULL;
1536 }
1537
1538 /**
1539  * Takes the next step in the execution, if possible.
1540  * @param curr The current step to take
1541  * @return Returns the next Thread to run, if any; NULL if this execution
1542  * should terminate
1543  */
1544 Thread * ModelExecution::take_step(ModelAction *curr)
1545 {
1546         Thread *curr_thrd = get_thread(curr);
1547         ASSERT(curr_thrd->get_state() == THREAD_READY);
1548
1549         ASSERT(check_action_enabled(curr));     /* May have side effects? */
1550         curr = check_current_action(curr);
1551         ASSERT(curr);
1552
1553         if (curr_thrd->is_blocked() || curr_thrd->is_complete())
1554                 scheduler->remove_thread(curr_thrd);
1555
1556         return action_select_next_thread(curr);
1557 }
1558
1559 Fuzzer * ModelExecution::getFuzzer() {
1560         return fuzzer;
1561 }