8 #include "snapshot-interface.h"
10 #include "clockvector.h"
11 #include "cyclegraph.h"
17 #define INITIAL_THREAD_ID 0
21 /** @brief Constructor */
22 ModelChecker::ModelChecker(struct model_params params) :
23 /* Initialize default scheduler */
25 scheduler(new Scheduler()),
27 num_feasible_executions(0),
29 earliest_diverge(NULL),
30 action_trace(new action_list_t()),
31 thread_map(new HashTable<int, Thread *, int>()),
32 obj_map(new HashTable<const void *, action_list_t, uintptr_t, 4>()),
33 lock_waiters_map(new HashTable<const void *, action_list_t, uintptr_t, 4>()),
34 obj_thrd_map(new HashTable<void *, std::vector<action_list_t>, uintptr_t, 4 >()),
35 promises(new std::vector< Promise *, SnapshotAlloc<Promise *> >()),
36 futurevalues(new std::vector< struct PendingFutureValue, SnapshotAlloc<struct PendingFutureValue> >()),
37 pending_rel_seqs(new std::vector< struct release_seq *, SnapshotAlloc<struct release_seq *> >()),
38 thrd_last_action(new std::vector< ModelAction *, SnapshotAlloc<ModelAction *> >(1)),
39 node_stack(new NodeStack()),
40 mo_graph(new CycleGraph()),
41 failed_promise(false),
42 too_many_reads(false),
44 bad_synchronization(false)
46 /* Allocate this "size" on the snapshotting heap */
47 priv = (struct model_snapshot_members *)calloc(1, sizeof(*priv));
48 /* First thread created will have id INITIAL_THREAD_ID */
49 priv->next_thread_id = INITIAL_THREAD_ID;
51 /* Initialize a model-checker thread, for special ModelActions */
52 model_thread = new Thread(get_next_id());
53 thread_map->put(id_to_int(model_thread->get_id()), model_thread);
56 /** @brief Destructor */
57 ModelChecker::~ModelChecker()
59 for (unsigned int i = 0; i < get_num_threads(); i++)
60 delete thread_map->get(i);
65 delete lock_waiters_map;
68 for (unsigned int i = 0; i < promises->size(); i++)
69 delete (*promises)[i];
72 delete pending_rel_seqs;
74 delete thrd_last_action;
81 * Restores user program to initial state and resets all model-checker data
84 void ModelChecker::reset_to_initial_state()
86 DEBUG("+++ Resetting to initial state +++\n");
87 node_stack->reset_execution();
88 failed_promise = false;
89 too_many_reads = false;
90 bad_synchronization = false;
92 snapshotObject->backTrackBeforeStep(0);
95 /** @return a thread ID for a new Thread */
96 thread_id_t ModelChecker::get_next_id()
98 return priv->next_thread_id++;
101 /** @return the number of user threads created during this execution */
102 unsigned int ModelChecker::get_num_threads()
104 return priv->next_thread_id;
107 /** @return The currently executing Thread. */
108 Thread * ModelChecker::get_current_thread()
110 return scheduler->get_current_thread();
113 /** @return a sequence number for a new ModelAction */
114 modelclock_t ModelChecker::get_next_seq_num()
116 return ++priv->used_sequence_numbers;
120 * @brief Choose the next thread to execute.
122 * This function chooses the next thread that should execute. It can force the
123 * adjacency of read/write portions of a RMW action, force THREAD_CREATE to be
124 * followed by a THREAD_START, or it can enforce execution replay/backtracking.
125 * The model-checker may have no preference regarding the next thread (i.e.,
126 * when exploring a new execution ordering), in which case this will return
128 * @param curr The current ModelAction. This action might guide the choice of
130 * @return The next thread to run. If the model-checker has no preference, NULL.
132 Thread * ModelChecker::get_next_thread(ModelAction *curr)
137 /* Do not split atomic actions. */
139 return thread_current();
140 /* The THREAD_CREATE action points to the created Thread */
141 else if (curr->get_type() == THREAD_CREATE)
142 return (Thread *)curr->get_location();
145 /* Have we completed exploring the preselected path? */
149 /* Else, we are trying to replay an execution */
150 ModelAction *next = node_stack->get_next()->get_action();
152 if (next == diverge) {
153 if (earliest_diverge == NULL || *diverge < *earliest_diverge)
154 earliest_diverge=diverge;
156 Node *nextnode = next->get_node();
157 Node *prevnode = nextnode->get_parent();
158 scheduler->update_sleep_set(prevnode);
160 /* Reached divergence point */
161 if (nextnode->increment_promise()) {
162 /* The next node will try to satisfy a different set of promises. */
163 tid = next->get_tid();
164 node_stack->pop_restofstack(2);
165 } else if (nextnode->increment_read_from()) {
166 /* The next node will read from a different value. */
167 tid = next->get_tid();
168 node_stack->pop_restofstack(2);
169 } else if (nextnode->increment_future_value()) {
170 /* The next node will try to read from a different future value. */
171 tid = next->get_tid();
172 node_stack->pop_restofstack(2);
173 } else if (nextnode->increment_relseq_break()) {
174 /* The next node will try to resolve a release sequence differently */
175 tid = next->get_tid();
176 node_stack->pop_restofstack(2);
178 /* Make a different thread execute for next step */
179 scheduler->add_sleep(thread_map->get(id_to_int(next->get_tid())));
180 tid = prevnode->get_next_backtrack();
181 /* Make sure the backtracked thread isn't sleeping. */
182 node_stack->pop_restofstack(1);
183 if (diverge==earliest_diverge) {
184 earliest_diverge=prevnode->get_action();
187 /* The correct sleep set is in the parent node. */
190 DEBUG("*** Divergence point ***\n");
194 tid = next->get_tid();
196 DEBUG("*** ModelChecker chose next thread = %d ***\n", id_to_int(tid));
197 ASSERT(tid != THREAD_ID_T_NONE);
198 return thread_map->get(id_to_int(tid));
202 * We need to know what the next actions of all threads in the sleep
203 * set will be. This method computes them and stores the actions at
204 * the corresponding thread object's pending action.
207 void ModelChecker::execute_sleep_set() {
208 for(unsigned int i=0;i<get_num_threads();i++) {
209 thread_id_t tid=int_to_id(i);
210 Thread *thr=get_thread(tid);
211 if ( scheduler->get_enabled(thr) == THREAD_SLEEP_SET ) {
212 thr->set_state(THREAD_RUNNING);
213 scheduler->next_thread(thr);
214 Thread::swap(&system_context, thr);
215 priv->current_action->set_sleep_flag();
216 thr->set_pending(priv->current_action);
219 priv->current_action = NULL;
222 void ModelChecker::wake_up_sleeping_actions(ModelAction * curr) {
223 for(unsigned int i=0;i<get_num_threads();i++) {
224 thread_id_t tid=int_to_id(i);
225 Thread *thr=get_thread(tid);
226 if ( scheduler->get_enabled(thr) == THREAD_SLEEP_SET ) {
227 ModelAction *pending_act=thr->get_pending();
228 if (pending_act->could_synchronize_with(curr)) {
229 //Remove this thread from sleep set
230 scheduler->remove_sleep(thr);
237 * Queries the model-checker for more executions to explore and, if one
238 * exists, resets the model-checker state to execute a new execution.
240 * @return If there are more executions to explore, return true. Otherwise,
243 bool ModelChecker::next_execution()
249 if (isfinalfeasible()) {
250 printf("Earliest divergence point since last feasible execution:\n");
251 if (earliest_diverge)
252 earliest_diverge->print();
254 printf("(Not set)\n");
256 earliest_diverge = NULL;
257 num_feasible_executions++;
260 DEBUG("Number of acquires waiting on pending release sequences: %zu\n",
261 pending_rel_seqs->size());
263 if (isfinalfeasible() || DBG_ENABLED())
266 if ((diverge = get_next_backtrack()) == NULL)
270 printf("Next execution will diverge at:\n");
274 reset_to_initial_state();
278 ModelAction * ModelChecker::get_last_conflict(ModelAction *act)
280 switch (act->get_type()) {
284 /* linear search: from most recent to oldest */
285 action_list_t *list = obj_map->get_safe_ptr(act->get_location());
286 action_list_t::reverse_iterator rit;
287 for (rit = list->rbegin(); rit != list->rend(); rit++) {
288 ModelAction *prev = *rit;
289 if (prev->could_synchronize_with(act))
295 case ATOMIC_TRYLOCK: {
296 /* linear search: from most recent to oldest */
297 action_list_t *list = obj_map->get_safe_ptr(act->get_location());
298 action_list_t::reverse_iterator rit;
299 for (rit = list->rbegin(); rit != list->rend(); rit++) {
300 ModelAction *prev = *rit;
301 if (act->is_conflicting_lock(prev))
306 case ATOMIC_UNLOCK: {
307 /* linear search: from most recent to oldest */
308 action_list_t *list = obj_map->get_safe_ptr(act->get_location());
309 action_list_t::reverse_iterator rit;
310 for (rit = list->rbegin(); rit != list->rend(); rit++) {
311 ModelAction *prev = *rit;
312 if (!act->same_thread(prev)&&prev->is_failed_trylock())
323 /** This method finds backtracking points where we should try to
324 * reorder the parameter ModelAction against.
326 * @param the ModelAction to find backtracking points for.
328 void ModelChecker::set_backtracking(ModelAction *act)
330 Thread *t = get_thread(act);
331 ModelAction * prev = get_last_conflict(act);
335 Node * node = prev->get_node()->get_parent();
337 int low_tid, high_tid;
338 if (node->is_enabled(t)) {
339 low_tid = id_to_int(act->get_tid());
340 high_tid = low_tid+1;
343 high_tid = get_num_threads();
346 for(int i = low_tid; i < high_tid; i++) {
347 thread_id_t tid = int_to_id(i);
349 /* Don't backtrack into a point where the thread is disabled or sleeping. */
350 if (node->get_enabled_array()[i]!=THREAD_ENABLED)
353 /* Check if this has been explored already */
354 if (node->has_been_explored(tid))
357 /* See if fairness allows */
358 if (model->params.fairwindow != 0 && !node->has_priority(tid)) {
360 for(int t=0;t<node->get_num_threads();t++) {
361 thread_id_t tother=int_to_id(t);
362 if (node->is_enabled(tother) && node->has_priority(tother)) {
370 /* Cache the latest backtracking point */
371 if (!priv->next_backtrack || *prev > *priv->next_backtrack)
372 priv->next_backtrack = prev;
374 /* If this is a new backtracking point, mark the tree */
375 if (!node->set_backtrack(tid))
377 DEBUG("Setting backtrack: conflict = %d, instead tid = %d\n",
378 id_to_int(prev->get_tid()),
379 id_to_int(t->get_id()));
388 * Returns last backtracking point. The model checker will explore a different
389 * path for this point in the next execution.
390 * @return The ModelAction at which the next execution should diverge.
392 ModelAction * ModelChecker::get_next_backtrack()
394 ModelAction *next = priv->next_backtrack;
395 priv->next_backtrack = NULL;
400 * Processes a read or rmw model action.
401 * @param curr is the read model action to process.
402 * @param second_part_of_rmw is boolean that is true is this is the second action of a rmw.
403 * @return True if processing this read updates the mo_graph.
405 bool ModelChecker::process_read(ModelAction *curr, bool second_part_of_rmw)
408 bool updated = false;
410 const ModelAction *reads_from = curr->get_node()->get_read_from();
411 if (reads_from != NULL) {
412 mo_graph->startChanges();
414 value = reads_from->get_value();
415 bool r_status = false;
417 if (!second_part_of_rmw) {
418 check_recency(curr, reads_from);
419 r_status = r_modification_order(curr, reads_from);
423 if (!second_part_of_rmw&&!isfeasible()&&(curr->get_node()->increment_read_from()||curr->get_node()->increment_future_value())) {
424 mo_graph->rollbackChanges();
425 too_many_reads = false;
429 curr->read_from(reads_from);
430 mo_graph->commitChanges();
431 mo_check_promises(curr->get_tid(), reads_from);
434 } else if (!second_part_of_rmw) {
435 /* Read from future value */
436 value = curr->get_node()->get_future_value();
437 modelclock_t expiration = curr->get_node()->get_future_value_expiration();
438 curr->read_from(NULL);
439 Promise *valuepromise = new Promise(curr, value, expiration);
440 promises->push_back(valuepromise);
442 get_thread(curr)->set_return_value(value);
448 * Processes a lock, trylock, or unlock model action. @param curr is
449 * the read model action to process.
451 * The try lock operation checks whether the lock is taken. If not,
452 * it falls to the normal lock operation case. If so, it returns
455 * The lock operation has already been checked that it is enabled, so
456 * it just grabs the lock and synchronizes with the previous unlock.
458 * The unlock operation has to re-enable all of the threads that are
459 * waiting on the lock.
461 * @return True if synchronization was updated; false otherwise
463 bool ModelChecker::process_mutex(ModelAction *curr) {
464 std::mutex *mutex = (std::mutex *)curr->get_location();
465 struct std::mutex_state *state = mutex->get_state();
466 switch (curr->get_type()) {
467 case ATOMIC_TRYLOCK: {
468 bool success = !state->islocked;
469 curr->set_try_lock(success);
471 get_thread(curr)->set_return_value(0);
474 get_thread(curr)->set_return_value(1);
476 //otherwise fall into the lock case
478 if (curr->get_cv()->getClock(state->alloc_tid) <= state->alloc_clock) {
479 printf("Lock access before initialization\n");
482 state->islocked = true;
483 ModelAction *unlock = get_last_unlock(curr);
484 //synchronize with the previous unlock statement
485 if (unlock != NULL) {
486 curr->synchronize_with(unlock);
491 case ATOMIC_UNLOCK: {
493 state->islocked = false;
494 //wake up the other threads
495 action_list_t *waiters = lock_waiters_map->get_safe_ptr(curr->get_location());
496 //activate all the waiting threads
497 for (action_list_t::iterator rit = waiters->begin(); rit != waiters->end(); rit++) {
498 scheduler->wake(get_thread(*rit));
510 * Process a write ModelAction
511 * @param curr The ModelAction to process
512 * @return True if the mo_graph was updated or promises were resolved
514 bool ModelChecker::process_write(ModelAction *curr)
516 bool updated_mod_order = w_modification_order(curr);
517 bool updated_promises = resolve_promises(curr);
519 if (promises->size() == 0) {
520 for (unsigned int i = 0; i < futurevalues->size(); i++) {
521 struct PendingFutureValue pfv = (*futurevalues)[i];
522 //Do more ambitious checks now that mo is more complete
523 if (mo_may_allow(pfv.writer, pfv.act)&&
524 pfv.act->get_node()->add_future_value(pfv.writer->get_value(), pfv.writer->get_seq_number()+params.maxfuturedelay) &&
525 (!priv->next_backtrack || *pfv.act > *priv->next_backtrack))
526 priv->next_backtrack = pfv.act;
528 futurevalues->resize(0);
531 mo_graph->commitChanges();
532 mo_check_promises(curr->get_tid(), curr);
534 get_thread(curr)->set_return_value(VALUE_NONE);
535 return updated_mod_order || updated_promises;
539 * @brief Process the current action for thread-related activity
541 * Performs current-action processing for a THREAD_* ModelAction. Proccesses
542 * may include setting Thread status, completing THREAD_FINISH/THREAD_JOIN
543 * synchronization, etc. This function is a no-op for non-THREAD actions
544 * (e.g., ATOMIC_{READ,WRITE,RMW,LOCK}, etc.)
546 * @param curr The current action
547 * @return True if synchronization was updated or a thread completed
549 bool ModelChecker::process_thread_action(ModelAction *curr)
551 bool updated = false;
553 switch (curr->get_type()) {
554 case THREAD_CREATE: {
555 Thread *th = (Thread *)curr->get_location();
556 th->set_creation(curr);
560 Thread *blocking = (Thread *)curr->get_location();
561 ModelAction *act = get_last_action(blocking->get_id());
562 curr->synchronize_with(act);
563 updated = true; /* trigger rel-seq checks */
566 case THREAD_FINISH: {
567 Thread *th = get_thread(curr);
568 while (!th->wait_list_empty()) {
569 ModelAction *act = th->pop_wait_list();
570 scheduler->wake(get_thread(act));
573 updated = true; /* trigger rel-seq checks */
577 check_promises(curr->get_tid(), NULL, curr->get_cv());
588 * @brief Process the current action for release sequence fixup activity
590 * Performs model-checker release sequence fixups for the current action,
591 * forcing a single pending release sequence to break (with a given, potential
592 * "loose" write) or to complete (i.e., synchronize). If a pending release
593 * sequence forms a complete release sequence, then we must perform the fixup
594 * synchronization, mo_graph additions, etc.
596 * @param curr The current action; must be a release sequence fixup action
597 * @param work_queue The work queue to which to add work items as they are
600 void ModelChecker::process_relseq_fixup(ModelAction *curr, work_queue_t *work_queue)
602 const ModelAction *write = curr->get_node()->get_relseq_break();
603 struct release_seq *sequence = pending_rel_seqs->back();
604 pending_rel_seqs->pop_back();
606 ModelAction *acquire = sequence->acquire;
607 const ModelAction *rf = sequence->rf;
608 const ModelAction *release = sequence->release;
612 ASSERT(release->same_thread(rf));
616 * @todo Forcing a synchronization requires that we set
617 * modification order constraints. For instance, we can't allow
618 * a fixup sequence in which two separate read-acquire
619 * operations read from the same sequence, where the first one
620 * synchronizes and the other doesn't. Essentially, we can't
621 * allow any writes to insert themselves between 'release' and
625 /* Must synchronize */
626 if (!acquire->synchronize_with(release)) {
627 set_bad_synchronization();
630 /* Re-check all pending release sequences */
631 work_queue->push_back(CheckRelSeqWorkEntry(NULL));
632 /* Re-check act for mo_graph edges */
633 work_queue->push_back(MOEdgeWorkEntry(acquire));
635 /* propagate synchronization to later actions */
636 action_list_t::reverse_iterator rit = action_trace->rbegin();
637 for (; (*rit) != acquire; rit++) {
638 ModelAction *propagate = *rit;
639 if (acquire->happens_before(propagate)) {
640 propagate->synchronize_with(acquire);
641 /* Re-check 'propagate' for mo_graph edges */
642 work_queue->push_back(MOEdgeWorkEntry(propagate));
646 /* Break release sequence with new edges:
647 * release --mo--> write --mo--> rf */
648 mo_graph->addEdge(release, write);
649 mo_graph->addEdge(write, rf);
652 /* See if we have realized a data race */
653 if (checkDataRaces())
658 * Initialize the current action by performing one or more of the following
659 * actions, as appropriate: merging RMWR and RMWC/RMW actions, stepping forward
660 * in the NodeStack, manipulating backtracking sets, allocating and
661 * initializing clock vectors, and computing the promises to fulfill.
663 * @param curr The current action, as passed from the user context; may be
664 * freed/invalidated after the execution of this function
665 * @return The current action, as processed by the ModelChecker. Is only the
666 * same as the parameter @a curr if this is a newly-explored action.
668 ModelAction * ModelChecker::initialize_curr_action(ModelAction *curr)
670 ModelAction *newcurr;
672 if (curr->is_rmwc() || curr->is_rmw()) {
673 newcurr = process_rmw(curr);
676 if (newcurr->is_rmw())
677 compute_promises(newcurr);
681 curr->set_seq_number(get_next_seq_num());
683 newcurr = node_stack->explore_action(curr, scheduler->get_enabled());
685 /* First restore type and order in case of RMW operation */
687 newcurr->copy_typeandorder(curr);
689 ASSERT(curr->get_location() == newcurr->get_location());
690 newcurr->copy_from_new(curr);
692 /* Discard duplicate ModelAction; use action from NodeStack */
695 /* Always compute new clock vector */
696 newcurr->create_cv(get_parent_action(newcurr->get_tid()));
700 /* Always compute new clock vector */
701 newcurr->create_cv(get_parent_action(newcurr->get_tid()));
703 * Perform one-time actions when pushing new ModelAction onto
706 if (newcurr->is_write())
707 compute_promises(newcurr);
708 else if (newcurr->is_relseq_fixup())
709 compute_relseq_breakwrites(newcurr);
715 * @brief Check whether a model action is enabled.
717 * Checks whether a lock or join operation would be successful (i.e., is the
718 * lock already locked, or is the joined thread already complete). If not, put
719 * the action in a waiter list.
721 * @param curr is the ModelAction to check whether it is enabled.
722 * @return a bool that indicates whether the action is enabled.
724 bool ModelChecker::check_action_enabled(ModelAction *curr) {
725 if (curr->is_lock()) {
726 std::mutex * lock = (std::mutex *)curr->get_location();
727 struct std::mutex_state * state = lock->get_state();
728 if (state->islocked) {
729 //Stick the action in the appropriate waiting queue
730 lock_waiters_map->get_safe_ptr(curr->get_location())->push_back(curr);
733 } else if (curr->get_type() == THREAD_JOIN) {
734 Thread *blocking = (Thread *)curr->get_location();
735 if (!blocking->is_complete()) {
736 blocking->push_wait_list(curr);
745 * This is the heart of the model checker routine. It performs model-checking
746 * actions corresponding to a given "current action." Among other processes, it
747 * calculates reads-from relationships, updates synchronization clock vectors,
748 * forms a memory_order constraints graph, and handles replay/backtrack
749 * execution when running permutations of previously-observed executions.
751 * @param curr The current action to process
752 * @return The next Thread that must be executed. May be NULL if ModelChecker
753 * makes no choice (e.g., according to replay execution, combining RMW actions,
756 Thread * ModelChecker::check_current_action(ModelAction *curr)
759 bool second_part_of_rmw = curr->is_rmwc() || curr->is_rmw();
761 if (!check_action_enabled(curr)) {
762 /* Make the execution look like we chose to run this action
763 * much later, when a lock/join can succeed */
764 get_current_thread()->set_pending(curr);
765 scheduler->sleep(get_current_thread());
766 return get_next_thread(NULL);
769 wake_up_sleeping_actions(curr);
771 ModelAction *newcurr = initialize_curr_action(curr);
774 /* Add the action to lists before any other model-checking tasks */
775 if (!second_part_of_rmw)
776 add_action_to_lists(newcurr);
778 /* Build may_read_from set for newly-created actions */
779 if (curr == newcurr && curr->is_read())
780 build_reads_from_past(curr);
783 /* Initialize work_queue with the "current action" work */
784 work_queue_t work_queue(1, CheckCurrWorkEntry(curr));
785 while (!work_queue.empty()) {
786 WorkQueueEntry work = work_queue.front();
787 work_queue.pop_front();
790 case WORK_CHECK_CURR_ACTION: {
791 ModelAction *act = work.action;
792 bool update = false; /* update this location's release seq's */
793 bool update_all = false; /* update all release seq's */
795 if (process_thread_action(curr))
798 if (act->is_read() && process_read(act, second_part_of_rmw))
801 if (act->is_write() && process_write(act))
804 if (act->is_mutex_op() && process_mutex(act))
807 if (act->is_relseq_fixup())
808 process_relseq_fixup(curr, &work_queue);
811 work_queue.push_back(CheckRelSeqWorkEntry(NULL));
813 work_queue.push_back(CheckRelSeqWorkEntry(act->get_location()));
816 case WORK_CHECK_RELEASE_SEQ:
817 resolve_release_sequences(work.location, &work_queue);
819 case WORK_CHECK_MO_EDGES: {
820 /** @todo Complete verification of work_queue */
821 ModelAction *act = work.action;
822 bool updated = false;
824 if (act->is_read()) {
825 const ModelAction *rf = act->get_reads_from();
826 if (rf != NULL && r_modification_order(act, rf))
829 if (act->is_write()) {
830 if (w_modification_order(act))
833 mo_graph->commitChanges();
836 work_queue.push_back(CheckRelSeqWorkEntry(act->get_location()));
845 check_curr_backtracking(curr);
846 set_backtracking(curr);
847 return get_next_thread(curr);
850 void ModelChecker::check_curr_backtracking(ModelAction * curr) {
851 Node *currnode = curr->get_node();
852 Node *parnode = currnode->get_parent();
854 if ((!parnode->backtrack_empty() ||
855 !currnode->read_from_empty() ||
856 !currnode->future_value_empty() ||
857 !currnode->promise_empty() ||
858 !currnode->relseq_break_empty())
859 && (!priv->next_backtrack ||
860 *curr > *priv->next_backtrack)) {
861 priv->next_backtrack = curr;
865 bool ModelChecker::promises_expired() {
866 for (unsigned int promise_index = 0; promise_index < promises->size(); promise_index++) {
867 Promise *promise = (*promises)[promise_index];
868 if (promise->get_expiration()<priv->used_sequence_numbers) {
875 /** @return whether the current partial trace must be a prefix of a
877 bool ModelChecker::isfeasibleprefix() {
878 return promises->size() == 0 && pending_rel_seqs->size() == 0;
881 /** @return whether the current partial trace is feasible. */
882 bool ModelChecker::isfeasible() {
883 if (DBG_ENABLED() && mo_graph->checkForRMWViolation())
884 DEBUG("Infeasible: RMW violation\n");
886 return !mo_graph->checkForRMWViolation() && isfeasibleotherthanRMW();
889 /** @return whether the current partial trace is feasible other than
890 * multiple RMW reading from the same store. */
891 bool ModelChecker::isfeasibleotherthanRMW() {
893 if (mo_graph->checkForCycles())
894 DEBUG("Infeasible: modification order cycles\n");
896 DEBUG("Infeasible: failed promise\n");
898 DEBUG("Infeasible: too many reads\n");
899 if (bad_synchronization)
900 DEBUG("Infeasible: bad synchronization ordering\n");
901 if (promises_expired())
902 DEBUG("Infeasible: promises expired\n");
904 return !mo_graph->checkForCycles() && !failed_promise && !too_many_reads && !bad_synchronization && !promises_expired();
907 /** Returns whether the current completed trace is feasible. */
908 bool ModelChecker::isfinalfeasible() {
909 if (DBG_ENABLED() && promises->size() != 0)
910 DEBUG("Infeasible: unrevolved promises\n");
912 return isfeasible() && promises->size() == 0;
915 /** Close out a RMWR by converting previous RMWR into a RMW or READ. */
916 ModelAction * ModelChecker::process_rmw(ModelAction *act) {
917 ModelAction *lastread = get_last_action(act->get_tid());
918 lastread->process_rmw(act);
919 if (act->is_rmw() && lastread->get_reads_from()!=NULL) {
920 mo_graph->addRMWEdge(lastread->get_reads_from(), lastread);
921 mo_graph->commitChanges();
927 * Checks whether a thread has read from the same write for too many times
928 * without seeing the effects of a later write.
931 * 1) there must a different write that we could read from that would satisfy the modification order,
932 * 2) we must have read from the same value in excess of maxreads times, and
933 * 3) that other write must have been in the reads_from set for maxreads times.
935 * If so, we decide that the execution is no longer feasible.
937 void ModelChecker::check_recency(ModelAction *curr, const ModelAction *rf) {
938 if (params.maxreads != 0) {
940 if (curr->get_node()->get_read_from_size() <= 1)
942 //Must make sure that execution is currently feasible... We could
943 //accidentally clear by rolling back
946 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(curr->get_location());
947 int tid = id_to_int(curr->get_tid());
950 if ((int)thrd_lists->size() <= tid)
952 action_list_t *list = &(*thrd_lists)[tid];
954 action_list_t::reverse_iterator rit = list->rbegin();
956 for (; (*rit) != curr; rit++)
958 /* go past curr now */
961 action_list_t::reverse_iterator ritcopy = rit;
962 //See if we have enough reads from the same value
964 for (; count < params.maxreads; rit++,count++) {
965 if (rit==list->rend())
967 ModelAction *act = *rit;
971 if (act->get_reads_from() != rf)
973 if (act->get_node()->get_read_from_size() <= 1)
976 for (int i = 0; i<curr->get_node()->get_read_from_size(); i++) {
978 const ModelAction * write = curr->get_node()->get_read_from_at(i);
980 //Need a different write
984 /* Test to see whether this is a feasible write to read from*/
985 mo_graph->startChanges();
986 r_modification_order(curr, write);
987 bool feasiblereadfrom = isfeasible();
988 mo_graph->rollbackChanges();
990 if (!feasiblereadfrom)
994 bool feasiblewrite = true;
995 //new we need to see if this write works for everyone
997 for (int loop = count; loop>0; loop--,rit++) {
998 ModelAction *act=*rit;
999 bool foundvalue = false;
1000 for (int j = 0; j<act->get_node()->get_read_from_size(); j++) {
1001 if (act->get_node()->get_read_from_at(i)==write) {
1007 feasiblewrite = false;
1011 if (feasiblewrite) {
1012 too_many_reads = true;
1020 * Updates the mo_graph with the constraints imposed from the current
1023 * Basic idea is the following: Go through each other thread and find
1024 * the lastest action that happened before our read. Two cases:
1026 * (1) The action is a write => that write must either occur before
1027 * the write we read from or be the write we read from.
1029 * (2) The action is a read => the write that that action read from
1030 * must occur before the write we read from or be the same write.
1032 * @param curr The current action. Must be a read.
1033 * @param rf The action that curr reads from. Must be a write.
1034 * @return True if modification order edges were added; false otherwise
1036 bool ModelChecker::r_modification_order(ModelAction *curr, const ModelAction *rf)
1038 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(curr->get_location());
1041 ASSERT(curr->is_read());
1043 /* Iterate over all threads */
1044 for (i = 0; i < thrd_lists->size(); i++) {
1045 /* Iterate over actions in thread, starting from most recent */
1046 action_list_t *list = &(*thrd_lists)[i];
1047 action_list_t::reverse_iterator rit;
1048 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1049 ModelAction *act = *rit;
1052 * Include at most one act per-thread that "happens
1053 * before" curr. Don't consider reflexively.
1055 if (act->happens_before(curr) && act != curr) {
1056 if (act->is_write()) {
1058 mo_graph->addEdge(act, rf);
1062 const ModelAction *prevreadfrom = act->get_reads_from();
1063 //if the previous read is unresolved, keep going...
1064 if (prevreadfrom == NULL)
1067 if (rf != prevreadfrom) {
1068 mo_graph->addEdge(prevreadfrom, rf);
1080 /** This method fixes up the modification order when we resolve a
1081 * promises. The basic problem is that actions that occur after the
1082 * read curr could not property add items to the modification order
1085 * So for each thread, we find the earliest item that happens after
1086 * the read curr. This is the item we have to fix up with additional
1087 * constraints. If that action is write, we add a MO edge between
1088 * the Action rf and that action. If the action is a read, we add a
1089 * MO edge between the Action rf, and whatever the read accessed.
1091 * @param curr is the read ModelAction that we are fixing up MO edges for.
1092 * @param rf is the write ModelAction that curr reads from.
1095 void ModelChecker::post_r_modification_order(ModelAction *curr, const ModelAction *rf)
1097 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(curr->get_location());
1099 ASSERT(curr->is_read());
1101 /* Iterate over all threads */
1102 for (i = 0; i < thrd_lists->size(); i++) {
1103 /* Iterate over actions in thread, starting from most recent */
1104 action_list_t *list = &(*thrd_lists)[i];
1105 action_list_t::reverse_iterator rit;
1106 ModelAction *lastact = NULL;
1108 /* Find last action that happens after curr that is either not curr or a rmw */
1109 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1110 ModelAction *act = *rit;
1111 if (curr->happens_before(act) && (curr != act || curr->is_rmw())) {
1117 /* Include at most one act per-thread that "happens before" curr */
1118 if (lastact != NULL) {
1119 if (lastact==curr) {
1120 //Case 1: The resolved read is a RMW, and we need to make sure
1121 //that the write portion of the RMW mod order after rf
1123 mo_graph->addEdge(rf, lastact);
1124 } else if (lastact->is_read()) {
1125 //Case 2: The resolved read is a normal read and the next
1126 //operation is a read, and we need to make sure the value read
1127 //is mod ordered after rf
1129 const ModelAction *postreadfrom = lastact->get_reads_from();
1130 if (postreadfrom != NULL&&rf != postreadfrom)
1131 mo_graph->addEdge(rf, postreadfrom);
1133 //Case 3: The resolved read is a normal read and the next
1134 //operation is a write, and we need to make sure that the
1135 //write is mod ordered after rf
1137 mo_graph->addEdge(rf, lastact);
1145 * Updates the mo_graph with the constraints imposed from the current write.
1147 * Basic idea is the following: Go through each other thread and find
1148 * the lastest action that happened before our write. Two cases:
1150 * (1) The action is a write => that write must occur before
1153 * (2) The action is a read => the write that that action read from
1154 * must occur before the current write.
1156 * This method also handles two other issues:
1158 * (I) Sequential Consistency: Making sure that if the current write is
1159 * seq_cst, that it occurs after the previous seq_cst write.
1161 * (II) Sending the write back to non-synchronizing reads.
1163 * @param curr The current action. Must be a write.
1164 * @return True if modification order edges were added; false otherwise
1166 bool ModelChecker::w_modification_order(ModelAction *curr)
1168 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(curr->get_location());
1171 ASSERT(curr->is_write());
1173 if (curr->is_seqcst()) {
1174 /* We have to at least see the last sequentially consistent write,
1175 so we are initialized. */
1176 ModelAction *last_seq_cst = get_last_seq_cst(curr);
1177 if (last_seq_cst != NULL) {
1178 mo_graph->addEdge(last_seq_cst, curr);
1183 /* Iterate over all threads */
1184 for (i = 0; i < thrd_lists->size(); i++) {
1185 /* Iterate over actions in thread, starting from most recent */
1186 action_list_t *list = &(*thrd_lists)[i];
1187 action_list_t::reverse_iterator rit;
1188 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1189 ModelAction *act = *rit;
1192 * 1) If RMW and it actually read from something, then we
1193 * already have all relevant edges, so just skip to next
1196 * 2) If RMW and it didn't read from anything, we should
1197 * whatever edge we can get to speed up convergence.
1199 * 3) If normal write, we need to look at earlier actions, so
1200 * continue processing list.
1202 if (curr->is_rmw()) {
1203 if (curr->get_reads_from()!=NULL)
1212 * Include at most one act per-thread that "happens
1215 if (act->happens_before(curr)) {
1217 * Note: if act is RMW, just add edge:
1219 * The following edge should be handled elsewhere:
1220 * readfrom(act) --mo--> act
1222 if (act->is_write())
1223 mo_graph->addEdge(act, curr);
1224 else if (act->is_read()) {
1225 //if previous read accessed a null, just keep going
1226 if (act->get_reads_from() == NULL)
1228 mo_graph->addEdge(act->get_reads_from(), curr);
1232 } else if (act->is_read() && !act->could_synchronize_with(curr) &&
1233 !act->same_thread(curr)) {
1234 /* We have an action that:
1235 (1) did not happen before us
1236 (2) is a read and we are a write
1237 (3) cannot synchronize with us
1238 (4) is in a different thread
1240 that read could potentially read from our write. Note that
1241 these checks are overly conservative at this point, we'll
1242 do more checks before actually removing the
1246 if (thin_air_constraint_may_allow(curr, act)) {
1248 (curr->is_rmw() && act->is_rmw() && curr->get_reads_from() == act->get_reads_from() && isfeasibleotherthanRMW())) {
1249 struct PendingFutureValue pfv = {curr,act};
1250 futurevalues->push_back(pfv);
1260 /** Arbitrary reads from the future are not allowed. Section 29.3
1261 * part 9 places some constraints. This method checks one result of constraint
1262 * constraint. Others require compiler support. */
1263 bool ModelChecker::thin_air_constraint_may_allow(const ModelAction * writer, const ModelAction *reader) {
1264 if (!writer->is_rmw())
1267 if (!reader->is_rmw())
1270 for (const ModelAction *search = writer->get_reads_from(); search != NULL; search = search->get_reads_from()) {
1271 if (search == reader)
1273 if (search->get_tid() == reader->get_tid() &&
1274 search->happens_before(reader))
1281 /** Arbitrary reads from the future are not allowed. Section 29.3
1282 * part 9 places some constraints. This method checks one result of constraint
1283 * constraint. Others require compiler support. */
1284 bool ModelChecker::mo_may_allow(const ModelAction * writer, const ModelAction *reader) {
1285 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(reader->get_location());
1287 //Get write that follows reader action
1288 action_list_t *list = &(*thrd_lists)[id_to_int(reader->get_tid())];
1289 action_list_t::reverse_iterator rit;
1290 ModelAction *first_write_after_read=NULL;
1292 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1293 ModelAction *act = *rit;
1296 if (act->is_write())
1297 first_write_after_read=act;
1300 if (first_write_after_read==NULL)
1303 return !mo_graph->checkReachable(first_write_after_read, writer);
1309 * Finds the head(s) of the release sequence(s) containing a given ModelAction.
1310 * The ModelAction under consideration is expected to be taking part in
1311 * release/acquire synchronization as an object of the "reads from" relation.
1312 * Note that this can only provide release sequence support for RMW chains
1313 * which do not read from the future, as those actions cannot be traced until
1314 * their "promise" is fulfilled. Similarly, we may not even establish the
1315 * presence of a release sequence with certainty, as some modification order
1316 * constraints may be decided further in the future. Thus, this function
1317 * "returns" two pieces of data: a pass-by-reference vector of @a release_heads
1318 * and a boolean representing certainty.
1320 * @param rf The action that might be part of a release sequence. Must be a
1322 * @param release_heads A pass-by-reference style return parameter. After
1323 * execution of this function, release_heads will contain the heads of all the
1324 * relevant release sequences, if any exists with certainty
1325 * @param pending A pass-by-reference style return parameter which is only used
1326 * when returning false (i.e., uncertain). Returns most information regarding
1327 * an uncertain release sequence, including any write operations that might
1328 * break the sequence.
1329 * @return true, if the ModelChecker is certain that release_heads is complete;
1332 bool ModelChecker::release_seq_heads(const ModelAction *rf,
1333 rel_heads_list_t *release_heads,
1334 struct release_seq *pending) const
1336 /* Only check for release sequences if there are no cycles */
1337 if (mo_graph->checkForCycles())
1341 ASSERT(rf->is_write());
1343 if (rf->is_release())
1344 release_heads->push_back(rf);
1346 break; /* End of RMW chain */
1348 /** @todo Need to be smarter here... In the linux lock
1349 * example, this will run to the beginning of the program for
1351 /** @todo The way to be smarter here is to keep going until 1
1352 * thread has a release preceded by an acquire and you've seen
1355 /* acq_rel RMW is a sufficient stopping condition */
1356 if (rf->is_acquire() && rf->is_release())
1357 return true; /* complete */
1359 rf = rf->get_reads_from();
1362 /* read from future: need to settle this later */
1364 return false; /* incomplete */
1367 if (rf->is_release())
1368 return true; /* complete */
1370 /* else relaxed write; check modification order for contiguous subsequence
1371 * -> rf must be same thread as release */
1372 int tid = id_to_int(rf->get_tid());
1373 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(rf->get_location());
1374 action_list_t *list = &(*thrd_lists)[tid];
1375 action_list_t::const_reverse_iterator rit;
1377 /* Find rf in the thread list */
1378 rit = std::find(list->rbegin(), list->rend(), rf);
1379 ASSERT(rit != list->rend());
1381 /* Find the last write/release */
1382 for (; rit != list->rend(); rit++)
1383 if ((*rit)->is_release())
1385 if (rit == list->rend()) {
1386 /* No write-release in this thread */
1387 return true; /* complete */
1389 ModelAction *release = *rit;
1391 ASSERT(rf->same_thread(release));
1393 pending->writes.clear();
1395 bool certain = true;
1396 for (unsigned int i = 0; i < thrd_lists->size(); i++) {
1397 if (id_to_int(rf->get_tid()) == (int)i)
1399 list = &(*thrd_lists)[i];
1401 /* Can we ensure no future writes from this thread may break
1402 * the release seq? */
1403 bool future_ordered = false;
1405 ModelAction *last = get_last_action(int_to_id(i));
1406 Thread *th = get_thread(int_to_id(i));
1407 if ((last && rf->happens_before(last)) ||
1408 !scheduler->is_enabled(th) ||
1410 future_ordered = true;
1412 ASSERT(!th->is_model_thread() || future_ordered);
1414 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1415 const ModelAction *act = *rit;
1416 /* Reach synchronization -> this thread is complete */
1417 if (act->happens_before(release))
1419 if (rf->happens_before(act)) {
1420 future_ordered = true;
1424 /* Only writes can break release sequences */
1425 if (!act->is_write())
1428 /* Check modification order */
1429 if (mo_graph->checkReachable(rf, act)) {
1430 /* rf --mo--> act */
1431 future_ordered = true;
1434 if (mo_graph->checkReachable(act, release))
1435 /* act --mo--> release */
1437 if (mo_graph->checkReachable(release, act) &&
1438 mo_graph->checkReachable(act, rf)) {
1439 /* release --mo-> act --mo--> rf */
1440 return true; /* complete */
1442 /* act may break release sequence */
1443 pending->writes.push_back(act);
1446 if (!future_ordered)
1447 certain = false; /* This thread is uncertain */
1451 release_heads->push_back(release);
1452 pending->writes.clear();
1454 pending->release = release;
1461 * A public interface for getting the release sequence head(s) with which a
1462 * given ModelAction must synchronize. This function only returns a non-empty
1463 * result when it can locate a release sequence head with certainty. Otherwise,
1464 * it may mark the internal state of the ModelChecker so that it will handle
1465 * the release sequence at a later time, causing @a act to update its
1466 * synchronization at some later point in execution.
1467 * @param act The 'acquire' action that may read from a release sequence
1468 * @param release_heads A pass-by-reference return parameter. Will be filled
1469 * with the head(s) of the release sequence(s), if they exists with certainty.
1470 * @see ModelChecker::release_seq_heads
1472 void ModelChecker::get_release_seq_heads(ModelAction *act, rel_heads_list_t *release_heads)
1474 const ModelAction *rf = act->get_reads_from();
1475 struct release_seq *sequence = (struct release_seq *)snapshot_calloc(1, sizeof(struct release_seq));
1476 sequence->acquire = act;
1478 if (!release_seq_heads(rf, release_heads, sequence)) {
1479 /* add act to 'lazy checking' list */
1480 pending_rel_seqs->push_back(sequence);
1482 snapshot_free(sequence);
1487 * Attempt to resolve all stashed operations that might synchronize with a
1488 * release sequence for a given location. This implements the "lazy" portion of
1489 * determining whether or not a release sequence was contiguous, since not all
1490 * modification order information is present at the time an action occurs.
1492 * @param location The location/object that should be checked for release
1493 * sequence resolutions. A NULL value means to check all locations.
1494 * @param work_queue The work queue to which to add work items as they are
1496 * @return True if any updates occurred (new synchronization, new mo_graph
1499 bool ModelChecker::resolve_release_sequences(void *location, work_queue_t *work_queue)
1501 bool updated = false;
1502 std::vector< struct release_seq *, SnapshotAlloc<struct release_seq *> >::iterator it = pending_rel_seqs->begin();
1503 while (it != pending_rel_seqs->end()) {
1504 struct release_seq *pending = *it;
1505 ModelAction *act = pending->acquire;
1507 /* Only resolve sequences on the given location, if provided */
1508 if (location && act->get_location() != location) {
1513 const ModelAction *rf = act->get_reads_from();
1514 rel_heads_list_t release_heads;
1516 complete = release_seq_heads(rf, &release_heads, pending);
1517 for (unsigned int i = 0; i < release_heads.size(); i++) {
1518 if (!act->has_synchronized_with(release_heads[i])) {
1519 if (act->synchronize_with(release_heads[i]))
1522 set_bad_synchronization();
1527 /* Re-check all pending release sequences */
1528 work_queue->push_back(CheckRelSeqWorkEntry(NULL));
1529 /* Re-check act for mo_graph edges */
1530 work_queue->push_back(MOEdgeWorkEntry(act));
1532 /* propagate synchronization to later actions */
1533 action_list_t::reverse_iterator rit = action_trace->rbegin();
1534 for (; (*rit) != act; rit++) {
1535 ModelAction *propagate = *rit;
1536 if (act->happens_before(propagate)) {
1537 propagate->synchronize_with(act);
1538 /* Re-check 'propagate' for mo_graph edges */
1539 work_queue->push_back(MOEdgeWorkEntry(propagate));
1544 it = pending_rel_seqs->erase(it);
1545 snapshot_free(pending);
1551 // If we resolved promises or data races, see if we have realized a data race.
1552 if (checkDataRaces()) {
1560 * Performs various bookkeeping operations for the current ModelAction. For
1561 * instance, adds action to the per-object, per-thread action vector and to the
1562 * action trace list of all thread actions.
1564 * @param act is the ModelAction to add.
1566 void ModelChecker::add_action_to_lists(ModelAction *act)
1568 int tid = id_to_int(act->get_tid());
1569 action_trace->push_back(act);
1571 obj_map->get_safe_ptr(act->get_location())->push_back(act);
1573 std::vector<action_list_t> *vec = obj_thrd_map->get_safe_ptr(act->get_location());
1574 if (tid >= (int)vec->size())
1575 vec->resize(priv->next_thread_id);
1576 (*vec)[tid].push_back(act);
1578 if ((int)thrd_last_action->size() <= tid)
1579 thrd_last_action->resize(get_num_threads());
1580 (*thrd_last_action)[tid] = act;
1584 * @brief Get the last action performed by a particular Thread
1585 * @param tid The thread ID of the Thread in question
1586 * @return The last action in the thread
1588 ModelAction * ModelChecker::get_last_action(thread_id_t tid) const
1590 int threadid = id_to_int(tid);
1591 if (threadid < (int)thrd_last_action->size())
1592 return (*thrd_last_action)[id_to_int(tid)];
1598 * Gets the last memory_order_seq_cst write (in the total global sequence)
1599 * performed on a particular object (i.e., memory location), not including the
1601 * @param curr The current ModelAction; also denotes the object location to
1603 * @return The last seq_cst write
1605 ModelAction * ModelChecker::get_last_seq_cst(ModelAction *curr) const
1607 void *location = curr->get_location();
1608 action_list_t *list = obj_map->get_safe_ptr(location);
1609 /* Find: max({i in dom(S) | seq_cst(t_i) && isWrite(t_i) && samevar(t_i, t)}) */
1610 action_list_t::reverse_iterator rit;
1611 for (rit = list->rbegin(); rit != list->rend(); rit++)
1612 if ((*rit)->is_write() && (*rit)->is_seqcst() && (*rit) != curr)
1618 * Gets the last unlock operation performed on a particular mutex (i.e., memory
1619 * location). This function identifies the mutex according to the current
1620 * action, which is presumed to perform on the same mutex.
1621 * @param curr The current ModelAction; also denotes the object location to
1623 * @return The last unlock operation
1625 ModelAction * ModelChecker::get_last_unlock(ModelAction *curr) const
1627 void *location = curr->get_location();
1628 action_list_t *list = obj_map->get_safe_ptr(location);
1629 /* Find: max({i in dom(S) | isUnlock(t_i) && samevar(t_i, t)}) */
1630 action_list_t::reverse_iterator rit;
1631 for (rit = list->rbegin(); rit != list->rend(); rit++)
1632 if ((*rit)->is_unlock())
1637 ModelAction * ModelChecker::get_parent_action(thread_id_t tid)
1639 ModelAction *parent = get_last_action(tid);
1641 parent = get_thread(tid)->get_creation();
1646 * Returns the clock vector for a given thread.
1647 * @param tid The thread whose clock vector we want
1648 * @return Desired clock vector
1650 ClockVector * ModelChecker::get_cv(thread_id_t tid)
1652 return get_parent_action(tid)->get_cv();
1656 * Resolve a set of Promises with a current write. The set is provided in the
1657 * Node corresponding to @a write.
1658 * @param write The ModelAction that is fulfilling Promises
1659 * @return True if promises were resolved; false otherwise
1661 bool ModelChecker::resolve_promises(ModelAction *write)
1663 bool resolved = false;
1664 std::vector< thread_id_t, ModelAlloc<thread_id_t> > threads_to_check;
1666 for (unsigned int i = 0, promise_index = 0; promise_index < promises->size(); i++) {
1667 Promise *promise = (*promises)[promise_index];
1668 if (write->get_node()->get_promise(i)) {
1669 ModelAction *read = promise->get_action();
1670 if (read->is_rmw()) {
1671 mo_graph->addRMWEdge(write, read);
1673 read->read_from(write);
1674 //First fix up the modification order for actions that happened
1676 r_modification_order(read, write);
1677 //Next fix up the modification order for actions that happened
1679 post_r_modification_order(read, write);
1680 //Make sure the promise's value matches the write's value
1681 ASSERT(promise->get_value() == write->get_value());
1684 promises->erase(promises->begin() + promise_index);
1685 threads_to_check.push_back(read->get_tid());
1692 //Check whether reading these writes has made threads unable to
1695 for(unsigned int i=0;i<threads_to_check.size();i++)
1696 mo_check_promises(threads_to_check[i], write);
1702 * Compute the set of promises that could potentially be satisfied by this
1703 * action. Note that the set computation actually appears in the Node, not in
1705 * @param curr The ModelAction that may satisfy promises
1707 void ModelChecker::compute_promises(ModelAction *curr)
1709 for (unsigned int i = 0; i < promises->size(); i++) {
1710 Promise *promise = (*promises)[i];
1711 const ModelAction *act = promise->get_action();
1712 if (!act->happens_before(curr) &&
1714 !act->could_synchronize_with(curr) &&
1715 !act->same_thread(curr) &&
1716 promise->get_value() == curr->get_value()) {
1717 curr->get_node()->set_promise(i);
1722 /** Checks promises in response to change in ClockVector Threads. */
1723 void ModelChecker::check_promises(thread_id_t tid, ClockVector *old_cv, ClockVector *merge_cv)
1725 for (unsigned int i = 0; i < promises->size(); i++) {
1726 Promise *promise = (*promises)[i];
1727 const ModelAction *act = promise->get_action();
1728 if ((old_cv == NULL || !old_cv->synchronized_since(act)) &&
1729 merge_cv->synchronized_since(act)) {
1730 if (promise->increment_threads(tid)) {
1731 //Promise has failed
1732 failed_promise = true;
1739 /** Checks promises in response to addition to modification order for threads.
1741 * pthread is the thread that performed the read that created the promise
1743 * pread is the read that created the promise
1745 * pwrite is either the first write to same location as pread by
1746 * pthread that is sequenced after pread or the value read by the
1747 * first read to the same lcoation as pread by pthread that is
1748 * sequenced after pread..
1750 * 1. If tid=pthread, then we check what other threads are reachable
1751 * through the mode order starting with pwrite. Those threads cannot
1752 * perform a write that will resolve the promise due to modification
1753 * order constraints.
1755 * 2. If the tid is not pthread, we check whether pwrite can reach the
1756 * action write through the modification order. If so, that thread
1757 * cannot perform a future write that will resolve the promise due to
1758 * modificatin order constraints.
1760 * @parem tid The thread that either read from the model action
1761 * write, or actually did the model action write.
1763 * @parem write The ModelAction representing the relevant write.
1766 void ModelChecker::mo_check_promises(thread_id_t tid, const ModelAction *write) {
1767 void * location = write->get_location();
1768 for (unsigned int i = 0; i < promises->size(); i++) {
1769 Promise *promise = (*promises)[i];
1770 const ModelAction *act = promise->get_action();
1772 //Is this promise on the same location?
1773 if ( act->get_location() != location )
1776 //same thread as the promise
1777 if ( act->get_tid()==tid ) {
1779 //do we have a pwrite for the promise, if not, set it
1780 if (promise->get_write() == NULL ) {
1781 promise->set_write(write);
1783 if (mo_graph->checkPromise(write, promise)) {
1784 failed_promise = true;
1789 //Don't do any lookups twice for the same thread
1790 if (promise->has_sync_thread(tid))
1793 if (mo_graph->checkReachable(promise->get_write(), write)) {
1794 if (promise->increment_threads(tid)) {
1795 failed_promise = true;
1803 * Compute the set of writes that may break the current pending release
1804 * sequence. This information is extracted from previou release sequence
1807 * @param curr The current ModelAction. Must be a release sequence fixup
1810 void ModelChecker::compute_relseq_breakwrites(ModelAction *curr)
1812 if (pending_rel_seqs->empty())
1815 struct release_seq *pending = pending_rel_seqs->back();
1816 for (unsigned int i = 0; i < pending->writes.size(); i++) {
1817 const ModelAction *write = pending->writes[i];
1818 curr->get_node()->add_relseq_break(write);
1821 /* NULL means don't break the sequence; just synchronize */
1822 curr->get_node()->add_relseq_break(NULL);
1826 * Build up an initial set of all past writes that this 'read' action may read
1827 * from. This set is determined by the clock vector's "happens before"
1829 * @param curr is the current ModelAction that we are exploring; it must be a
1832 void ModelChecker::build_reads_from_past(ModelAction *curr)
1834 std::vector<action_list_t> *thrd_lists = obj_thrd_map->get_safe_ptr(curr->get_location());
1836 ASSERT(curr->is_read());
1838 ModelAction *last_seq_cst = NULL;
1840 /* Track whether this object has been initialized */
1841 bool initialized = false;
1843 if (curr->is_seqcst()) {
1844 last_seq_cst = get_last_seq_cst(curr);
1845 /* We have to at least see the last sequentially consistent write,
1846 so we are initialized. */
1847 if (last_seq_cst != NULL)
1851 /* Iterate over all threads */
1852 for (i = 0; i < thrd_lists->size(); i++) {
1853 /* Iterate over actions in thread, starting from most recent */
1854 action_list_t *list = &(*thrd_lists)[i];
1855 action_list_t::reverse_iterator rit;
1856 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1857 ModelAction *act = *rit;
1859 /* Only consider 'write' actions */
1860 if (!act->is_write() || act == curr)
1863 /* Don't consider more than one seq_cst write if we are a seq_cst read. */
1864 if (!curr->is_seqcst() || (!act->is_seqcst() && (last_seq_cst == NULL || !act->happens_before(last_seq_cst))) || act == last_seq_cst) {
1865 DEBUG("Adding action to may_read_from:\n");
1866 if (DBG_ENABLED()) {
1871 if (curr->get_sleep_flag()) {
1872 if (sleep_can_read_from(curr, act))
1873 curr->get_node()->add_read_from(act);
1875 curr->get_node()->add_read_from(act);
1878 /* Include at most one act per-thread that "happens before" curr */
1879 if (act->happens_before(curr)) {
1887 /** @todo Need a more informative way of reporting errors. */
1888 printf("ERROR: may read from uninitialized atomic\n");
1891 if (DBG_ENABLED() || !initialized) {
1892 printf("Reached read action:\n");
1894 printf("Printing may_read_from\n");
1895 curr->get_node()->print_may_read_from();
1896 printf("End printing may_read_from\n");
1899 ASSERT(initialized);
1902 bool ModelChecker::sleep_can_read_from(ModelAction * curr, const ModelAction *write) {
1904 Node *prevnode=write->get_node()->get_parent();
1905 bool thread_sleep=prevnode->get_enabled_array()[id_to_int(curr->get_tid())]==THREAD_SLEEP_SET;
1906 if (write->is_release()&&thread_sleep)
1908 if (!write->is_rmw()) {
1911 if (write->get_reads_from()==NULL)
1913 write=write->get_reads_from();
1917 static void print_list(action_list_t *list)
1919 action_list_t::iterator it;
1921 printf("---------------------------------------------------------------------\n");
1923 unsigned int hash=0;
1925 for (it = list->begin(); it != list->end(); it++) {
1927 hash=hash^(hash<<3)^((*it)->hash());
1929 printf("HASH %u\n", hash);
1930 printf("---------------------------------------------------------------------\n");
1933 #if SUPPORT_MOD_ORDER_DUMP
1934 void ModelChecker::dumpGraph(char *filename) {
1936 sprintf(buffer, "%s.dot",filename);
1937 FILE *file=fopen(buffer, "w");
1938 fprintf(file, "digraph %s {\n",filename);
1939 mo_graph->dumpNodes(file);
1940 ModelAction ** thread_array=(ModelAction **)model_calloc(1, sizeof(ModelAction *)*get_num_threads());
1942 for (action_list_t::iterator it = action_trace->begin(); it != action_trace->end(); it++) {
1943 ModelAction *action=*it;
1944 if (action->is_read()) {
1945 fprintf(file, "N%u [label=\"%u, T%u\"];\n", action->get_seq_number(),action->get_seq_number(), action->get_tid());
1946 if (action->get_reads_from()!=NULL)
1947 fprintf(file, "N%u -> N%u[label=\"rf\", color=red];\n", action->get_seq_number(), action->get_reads_from()->get_seq_number());
1949 if (thread_array[action->get_tid()] != NULL) {
1950 fprintf(file, "N%u -> N%u[label=\"sb\", color=blue];\n", thread_array[action->get_tid()]->get_seq_number(), action->get_seq_number());
1953 thread_array[action->get_tid()]=action;
1955 fprintf(file,"}\n");
1956 model_free(thread_array);
1961 void ModelChecker::print_summary()
1964 printf("Number of executions: %d\n", num_executions);
1965 printf("Number of feasible executions: %d\n", num_feasible_executions);
1966 printf("Total nodes created: %d\n", node_stack->get_total_nodes());
1968 #if SUPPORT_MOD_ORDER_DUMP
1970 char buffername[100];
1971 sprintf(buffername, "exec%04u", num_executions);
1972 mo_graph->dumpGraphToFile(buffername);
1973 sprintf(buffername, "graph%04u", num_executions);
1974 dumpGraph(buffername);
1977 if (!isfinalfeasible())
1978 printf("INFEASIBLE EXECUTION!\n");
1979 print_list(action_trace);
1984 * Add a Thread to the system for the first time. Should only be called once
1986 * @param t The Thread to add
1988 void ModelChecker::add_thread(Thread *t)
1990 thread_map->put(id_to_int(t->get_id()), t);
1991 scheduler->add_thread(t);
1995 * Removes a thread from the scheduler.
1996 * @param the thread to remove.
1998 void ModelChecker::remove_thread(Thread *t)
2000 scheduler->remove_thread(t);
2004 * @brief Get a Thread reference by its ID
2005 * @param tid The Thread's ID
2006 * @return A Thread reference
2008 Thread * ModelChecker::get_thread(thread_id_t tid) const
2010 return thread_map->get(id_to_int(tid));
2014 * @brief Get a reference to the Thread in which a ModelAction was executed
2015 * @param act The ModelAction
2016 * @return A Thread reference
2018 Thread * ModelChecker::get_thread(ModelAction *act) const
2020 return get_thread(act->get_tid());
2024 * Switch from a user-context to the "master thread" context (a.k.a. system
2025 * context). This switch is made with the intention of exploring a particular
2026 * model-checking action (described by a ModelAction object). Must be called
2027 * from a user-thread context.
2029 * @param act The current action that will be explored. May be NULL only if
2030 * trace is exiting via an assertion (see ModelChecker::set_assert and
2031 * ModelChecker::has_asserted).
2032 * @return Return status from the 'swap' call (i.e., success/fail, 0/-1)
2034 int ModelChecker::switch_to_master(ModelAction *act)
2037 Thread *old = thread_current();
2038 set_current_action(act);
2039 old->set_state(THREAD_READY);
2040 return Thread::swap(old, &system_context);
2044 * Takes the next step in the execution, if possible.
2045 * @return Returns true (success) if a step was taken and false otherwise.
2047 bool ModelChecker::take_step() {
2051 Thread *curr = priv->current_action ? get_thread(priv->current_action) : NULL;
2053 if (curr->get_state() == THREAD_READY) {
2054 ASSERT(priv->current_action);
2056 priv->nextThread = check_current_action(priv->current_action);
2057 priv->current_action = NULL;
2059 if (curr->is_blocked() || curr->is_complete())
2060 scheduler->remove_thread(curr);
2065 Thread *next = scheduler->next_thread(priv->nextThread);
2067 /* Infeasible -> don't take any more steps */
2071 DEBUG("(%d, %d)\n", curr ? id_to_int(curr->get_id()) : -1,
2072 next ? id_to_int(next->get_id()) : -1);
2075 * Launch end-of-execution release sequence fixups only when there are:
2077 * (1) no more user threads to run (or when execution replay chooses
2078 * the 'model_thread')
2079 * (2) pending release sequences
2080 * (3) pending assertions (i.e., data races)
2081 * (4) no pending promises
2083 if (!pending_rel_seqs->empty() && (!next || next->is_model_thread()) &&
2084 isfinalfeasible() && !unrealizedraces.empty()) {
2085 printf("*** WARNING: release sequence fixup action (%zu pending release seuqences) ***\n",
2086 pending_rel_seqs->size());
2087 ModelAction *fixup = new ModelAction(MODEL_FIXUP_RELSEQ,
2088 std::memory_order_seq_cst, NULL, VALUE_NONE,
2090 set_current_action(fixup);
2094 /* next == NULL -> don't take any more steps */
2098 next->set_state(THREAD_RUNNING);
2100 if (next->get_pending() != NULL) {
2101 /* restart a pending action */
2102 set_current_action(next->get_pending());
2103 next->set_pending(NULL);
2104 next->set_state(THREAD_READY);
2108 /* Return false only if swap fails with an error */
2109 return (Thread::swap(&system_context, next) == 0);
2112 /** Runs the current execution until threre are no more steps to take. */
2113 void ModelChecker::finish_execution() {
2116 while (take_step());