9 #include "snapshot-interface.h"
11 #include "clockvector.h"
12 #include "cyclegraph.h"
15 #include "threads-model.h"
18 #define INITIAL_THREAD_ID 0
23 bug_message(const char *str) {
24 const char *fmt = " [BUG] %s\n";
25 msg = (char *)snapshot_malloc(strlen(fmt) + strlen(str));
26 sprintf(msg, fmt, str);
28 ~bug_message() { if (msg) snapshot_free(msg); }
31 void print() { model_print("%s", msg); }
37 * Structure for holding small ModelChecker members that should be snapshotted
39 struct model_snapshot_members {
40 model_snapshot_members() :
42 /* First thread created will have id INITIAL_THREAD_ID */
43 next_thread_id(INITIAL_THREAD_ID),
44 used_sequence_numbers(0),
49 failed_promise(false),
50 too_many_reads(false),
51 bad_synchronization(false),
55 ~model_snapshot_members() {
56 for (unsigned int i = 0; i < bugs.size(); i++)
61 ModelAction *current_action;
62 unsigned int next_thread_id;
63 modelclock_t used_sequence_numbers;
65 ModelAction *next_backtrack;
66 std::vector< bug_message *, SnapshotAlloc<bug_message *> > bugs;
67 struct execution_stats stats;
70 /** @brief Incorrectly-ordered synchronization was made */
71 bool bad_synchronization;
77 /** @brief Constructor */
78 ModelChecker::ModelChecker(struct model_params params) :
79 /* Initialize default scheduler */
81 scheduler(new Scheduler()),
83 earliest_diverge(NULL),
84 action_trace(new action_list_t()),
85 thread_map(new HashTable<int, Thread *, int>()),
86 obj_map(new HashTable<const void *, action_list_t *, uintptr_t, 4>()),
87 lock_waiters_map(new HashTable<const void *, action_list_t *, uintptr_t, 4>()),
88 condvar_waiters_map(new HashTable<const void *, action_list_t *, uintptr_t, 4>()),
89 obj_thrd_map(new HashTable<void *, std::vector<action_list_t> *, uintptr_t, 4 >()),
90 promises(new std::vector< Promise *, SnapshotAlloc<Promise *> >()),
91 futurevalues(new std::vector< struct PendingFutureValue, SnapshotAlloc<struct PendingFutureValue> >()),
92 pending_rel_seqs(new std::vector< struct release_seq *, SnapshotAlloc<struct release_seq *> >()),
93 thrd_last_action(new std::vector< ModelAction *, SnapshotAlloc<ModelAction *> >(1)),
94 thrd_last_fence_release(new std::vector< ModelAction *, SnapshotAlloc<ModelAction *> >()),
95 node_stack(new NodeStack()),
96 priv(new struct model_snapshot_members()),
97 mo_graph(new CycleGraph())
99 /* Initialize a model-checker thread, for special ModelActions */
100 model_thread = new Thread(get_next_id());
101 thread_map->put(id_to_int(model_thread->get_id()), model_thread);
104 /** @brief Destructor */
105 ModelChecker::~ModelChecker()
107 for (unsigned int i = 0; i < get_num_threads(); i++)
108 delete thread_map->get(i);
113 delete lock_waiters_map;
114 delete condvar_waiters_map;
117 for (unsigned int i = 0; i < promises->size(); i++)
118 delete (*promises)[i];
121 delete pending_rel_seqs;
123 delete thrd_last_action;
124 delete thrd_last_fence_release;
131 static action_list_t * get_safe_ptr_action(HashTable<const void *, action_list_t *, uintptr_t, 4> * hash, void * ptr) {
132 action_list_t * tmp=hash->get(ptr);
134 tmp=new action_list_t();
140 static std::vector<action_list_t> * get_safe_ptr_vect_action(HashTable<void *, std::vector<action_list_t> *, uintptr_t, 4> * hash, void * ptr) {
141 std::vector<action_list_t> * tmp=hash->get(ptr);
143 tmp=new std::vector<action_list_t>();
150 * Restores user program to initial state and resets all model-checker data
153 void ModelChecker::reset_to_initial_state()
155 DEBUG("+++ Resetting to initial state +++\n");
156 node_stack->reset_execution();
158 /* Print all model-checker output before rollback */
161 snapshotObject->backTrackBeforeStep(0);
164 /** @return a thread ID for a new Thread */
165 thread_id_t ModelChecker::get_next_id()
167 return priv->next_thread_id++;
170 /** @return the number of user threads created during this execution */
171 unsigned int ModelChecker::get_num_threads() const
173 return priv->next_thread_id;
176 /** @return The currently executing Thread. */
177 Thread * ModelChecker::get_current_thread() const
179 return scheduler->get_current_thread();
182 /** @return a sequence number for a new ModelAction */
183 modelclock_t ModelChecker::get_next_seq_num()
185 return ++priv->used_sequence_numbers;
188 Node * ModelChecker::get_curr_node() const
190 return node_stack->get_head();
194 * @brief Choose the next thread to execute.
196 * This function chooses the next thread that should execute. It can force the
197 * adjacency of read/write portions of a RMW action, force THREAD_CREATE to be
198 * followed by a THREAD_START, or it can enforce execution replay/backtracking.
199 * The model-checker may have no preference regarding the next thread (i.e.,
200 * when exploring a new execution ordering), in which case this will return
202 * @param curr The current ModelAction. This action might guide the choice of
204 * @return The next thread to run. If the model-checker has no preference, NULL.
206 Thread * ModelChecker::get_next_thread(ModelAction *curr)
211 /* Do not split atomic actions. */
213 return thread_current();
214 /* The THREAD_CREATE action points to the created Thread */
215 else if (curr->get_type() == THREAD_CREATE)
216 return (Thread *)curr->get_location();
219 /* Have we completed exploring the preselected path? */
223 /* Else, we are trying to replay an execution */
224 ModelAction *next = node_stack->get_next()->get_action();
226 if (next == diverge) {
227 if (earliest_diverge == NULL || *diverge < *earliest_diverge)
228 earliest_diverge=diverge;
230 Node *nextnode = next->get_node();
231 Node *prevnode = nextnode->get_parent();
232 scheduler->update_sleep_set(prevnode);
234 /* Reached divergence point */
235 if (nextnode->increment_misc()) {
236 /* The next node will try to satisfy a different misc_index values. */
237 tid = next->get_tid();
238 node_stack->pop_restofstack(2);
239 } else if (nextnode->increment_promise()) {
240 /* The next node will try to satisfy a different set of promises. */
241 tid = next->get_tid();
242 node_stack->pop_restofstack(2);
243 } else if (nextnode->increment_read_from()) {
244 /* The next node will read from a different value. */
245 tid = next->get_tid();
246 node_stack->pop_restofstack(2);
247 } else if (nextnode->increment_future_value()) {
248 /* The next node will try to read from a different future value. */
249 tid = next->get_tid();
250 node_stack->pop_restofstack(2);
251 } else if (nextnode->increment_relseq_break()) {
252 /* The next node will try to resolve a release sequence differently */
253 tid = next->get_tid();
254 node_stack->pop_restofstack(2);
256 /* Make a different thread execute for next step */
257 scheduler->add_sleep(thread_map->get(id_to_int(next->get_tid())));
258 tid = prevnode->get_next_backtrack();
259 /* Make sure the backtracked thread isn't sleeping. */
260 node_stack->pop_restofstack(1);
261 if (diverge==earliest_diverge) {
262 earliest_diverge=prevnode->get_action();
265 /* The correct sleep set is in the parent node. */
268 DEBUG("*** Divergence point ***\n");
272 tid = next->get_tid();
274 DEBUG("*** ModelChecker chose next thread = %d ***\n", id_to_int(tid));
275 ASSERT(tid != THREAD_ID_T_NONE);
276 return thread_map->get(id_to_int(tid));
280 * We need to know what the next actions of all threads in the sleep
281 * set will be. This method computes them and stores the actions at
282 * the corresponding thread object's pending action.
285 void ModelChecker::execute_sleep_set() {
286 for(unsigned int i=0;i<get_num_threads();i++) {
287 thread_id_t tid=int_to_id(i);
288 Thread *thr=get_thread(tid);
289 if ( scheduler->get_enabled(thr) == THREAD_SLEEP_SET &&
290 thr->get_pending() == NULL ) {
291 thr->set_state(THREAD_RUNNING);
292 scheduler->next_thread(thr);
293 Thread::swap(&system_context, thr);
294 priv->current_action->set_sleep_flag();
295 thr->set_pending(priv->current_action);
298 priv->current_action = NULL;
301 void ModelChecker::wake_up_sleeping_actions(ModelAction * curr) {
302 for(unsigned int i=0;i<get_num_threads();i++) {
303 thread_id_t tid=int_to_id(i);
304 Thread *thr=get_thread(tid);
305 if ( scheduler->get_enabled(thr) == THREAD_SLEEP_SET ) {
306 ModelAction *pending_act=thr->get_pending();
307 if ((!curr->is_rmwr())&&pending_act->could_synchronize_with(curr)) {
308 //Remove this thread from sleep set
309 scheduler->remove_sleep(thr);
315 /** @brief Alert the model-checker that an incorrectly-ordered
316 * synchronization was made */
317 void ModelChecker::set_bad_synchronization()
319 priv->bad_synchronization = true;
322 bool ModelChecker::has_asserted() const
324 return priv->asserted;
327 void ModelChecker::set_assert()
329 priv->asserted = true;
333 * Check if we are in a deadlock. Should only be called at the end of an
334 * execution, although it should not give false positives in the middle of an
335 * execution (there should be some ENABLED thread).
337 * @return True if program is in a deadlock; false otherwise
339 bool ModelChecker::is_deadlocked() const
341 bool blocking_threads = false;
342 for (unsigned int i = 0; i < get_num_threads(); i++) {
343 thread_id_t tid = int_to_id(i);
346 Thread *t = get_thread(tid);
347 if (!t->is_model_thread() && t->get_pending())
348 blocking_threads = true;
350 return blocking_threads;
354 * Check if this is a complete execution. That is, have all thread completed
355 * execution (rather than exiting because sleep sets have forced a redundant
358 * @return True if the execution is complete.
360 bool ModelChecker::is_complete_execution() const
362 for (unsigned int i = 0; i < get_num_threads(); i++)
363 if (is_enabled(int_to_id(i)))
369 * @brief Assert a bug in the executing program.
371 * Use this function to assert any sort of bug in the user program. If the
372 * current trace is feasible (actually, a prefix of some feasible execution),
373 * then this execution will be aborted, printing the appropriate message. If
374 * the current trace is not yet feasible, the error message will be stashed and
375 * printed if the execution ever becomes feasible.
377 * @param msg Descriptive message for the bug (do not include newline char)
378 * @return True if bug is immediately-feasible
380 bool ModelChecker::assert_bug(const char *msg)
382 priv->bugs.push_back(new bug_message(msg));
384 if (isfeasibleprefix()) {
392 * @brief Assert a bug in the executing program, asserted by a user thread
393 * @see ModelChecker::assert_bug
394 * @param msg Descriptive message for the bug (do not include newline char)
396 void ModelChecker::assert_user_bug(const char *msg)
398 /* If feasible bug, bail out now */
400 switch_to_master(NULL);
403 /** @return True, if any bugs have been reported for this execution */
404 bool ModelChecker::have_bug_reports() const
406 return priv->bugs.size() != 0;
409 /** @brief Print bug report listing for this execution (if any bugs exist) */
410 void ModelChecker::print_bugs() const
412 if (have_bug_reports()) {
413 model_print("Bug report: %zu bug%s detected\n",
415 priv->bugs.size() > 1 ? "s" : "");
416 for (unsigned int i = 0; i < priv->bugs.size(); i++)
417 priv->bugs[i]->print();
422 * @brief Record end-of-execution stats
424 * Must be run when exiting an execution. Records various stats.
425 * @see struct execution_stats
427 void ModelChecker::record_stats()
430 if (!isfeasibleprefix())
431 stats.num_infeasible++;
432 else if (have_bug_reports())
433 stats.num_buggy_executions++;
434 else if (is_complete_execution())
435 stats.num_complete++;
437 stats.num_redundant++;
440 /** @brief Print execution stats */
441 void ModelChecker::print_stats() const
443 model_print("Number of complete, bug-free executions: %d\n", stats.num_complete);
444 model_print("Number of redundant executions: %d\n", stats.num_redundant);
445 model_print("Number of buggy executions: %d\n", stats.num_buggy_executions);
446 model_print("Number of infeasible executions: %d\n", stats.num_infeasible);
447 model_print("Total executions: %d\n", stats.num_total);
448 model_print("Total nodes created: %d\n", node_stack->get_total_nodes());
452 * @brief End-of-exeuction print
453 * @param printbugs Should any existing bugs be printed?
455 void ModelChecker::print_execution(bool printbugs) const
457 print_program_output();
459 if (DBG_ENABLED() || params.verbose) {
460 model_print("Earliest divergence point since last feasible execution:\n");
461 if (earliest_diverge)
462 earliest_diverge->print();
464 model_print("(Not set)\n");
470 /* Don't print invalid bugs */
479 * Queries the model-checker for more executions to explore and, if one
480 * exists, resets the model-checker state to execute a new execution.
482 * @return If there are more executions to explore, return true. Otherwise,
485 bool ModelChecker::next_execution()
488 /* Is this execution a feasible execution that's worth bug-checking? */
489 bool complete = isfeasibleprefix() && (is_complete_execution() ||
492 /* End-of-execution bug checks */
495 assert_bug("Deadlock detected");
503 if (DBG_ENABLED() || params.verbose || have_bug_reports())
504 print_execution(complete);
506 clear_program_output();
509 earliest_diverge = NULL;
511 if ((diverge = get_next_backtrack()) == NULL)
515 model_print("Next execution will diverge at:\n");
519 reset_to_initial_state();
523 ModelAction * ModelChecker::get_last_conflict(ModelAction *act)
525 switch (act->get_type()) {
530 /* Optimization: relaxed operations don't need backtracking */
531 if (act->is_relaxed())
533 /* linear search: from most recent to oldest */
534 action_list_t *list = get_safe_ptr_action(obj_map, act->get_location());
535 action_list_t::reverse_iterator rit;
536 for (rit = list->rbegin(); rit != list->rend(); rit++) {
537 ModelAction *prev = *rit;
538 if (prev->could_synchronize_with(act))
544 case ATOMIC_TRYLOCK: {
545 /* linear search: from most recent to oldest */
546 action_list_t *list = get_safe_ptr_action(obj_map, act->get_location());
547 action_list_t::reverse_iterator rit;
548 for (rit = list->rbegin(); rit != list->rend(); rit++) {
549 ModelAction *prev = *rit;
550 if (act->is_conflicting_lock(prev))
555 case ATOMIC_UNLOCK: {
556 /* linear search: from most recent to oldest */
557 action_list_t *list = get_safe_ptr_action(obj_map, act->get_location());
558 action_list_t::reverse_iterator rit;
559 for (rit = list->rbegin(); rit != list->rend(); rit++) {
560 ModelAction *prev = *rit;
561 if (!act->same_thread(prev)&&prev->is_failed_trylock())
567 /* linear search: from most recent to oldest */
568 action_list_t *list = get_safe_ptr_action(obj_map, act->get_location());
569 action_list_t::reverse_iterator rit;
570 for (rit = list->rbegin(); rit != list->rend(); rit++) {
571 ModelAction *prev = *rit;
572 if (!act->same_thread(prev)&&prev->is_failed_trylock())
574 if (!act->same_thread(prev)&&prev->is_notify())
580 case ATOMIC_NOTIFY_ALL:
581 case ATOMIC_NOTIFY_ONE: {
582 /* linear search: from most recent to oldest */
583 action_list_t *list = get_safe_ptr_action(obj_map, act->get_location());
584 action_list_t::reverse_iterator rit;
585 for (rit = list->rbegin(); rit != list->rend(); rit++) {
586 ModelAction *prev = *rit;
587 if (!act->same_thread(prev)&&prev->is_wait())
598 /** This method finds backtracking points where we should try to
599 * reorder the parameter ModelAction against.
601 * @param the ModelAction to find backtracking points for.
603 void ModelChecker::set_backtracking(ModelAction *act)
605 Thread *t = get_thread(act);
606 ModelAction * prev = get_last_conflict(act);
610 Node * node = prev->get_node()->get_parent();
612 int low_tid, high_tid;
613 if (node->is_enabled(t)) {
614 low_tid = id_to_int(act->get_tid());
615 high_tid = low_tid+1;
618 high_tid = get_num_threads();
621 for(int i = low_tid; i < high_tid; i++) {
622 thread_id_t tid = int_to_id(i);
624 /* Make sure this thread can be enabled here. */
625 if (i >= node->get_num_threads())
628 /* Don't backtrack into a point where the thread is disabled or sleeping. */
629 if (node->enabled_status(tid)!=THREAD_ENABLED)
632 /* Check if this has been explored already */
633 if (node->has_been_explored(tid))
636 /* See if fairness allows */
637 if (model->params.fairwindow != 0 && !node->has_priority(tid)) {
639 for(int t=0;t<node->get_num_threads();t++) {
640 thread_id_t tother=int_to_id(t);
641 if (node->is_enabled(tother) && node->has_priority(tother)) {
649 /* Cache the latest backtracking point */
650 if (!priv->next_backtrack || *prev > *priv->next_backtrack)
651 priv->next_backtrack = prev;
653 /* If this is a new backtracking point, mark the tree */
654 if (!node->set_backtrack(tid))
656 DEBUG("Setting backtrack: conflict = %d, instead tid = %d\n",
657 id_to_int(prev->get_tid()),
658 id_to_int(t->get_id()));
667 * Returns last backtracking point. The model checker will explore a different
668 * path for this point in the next execution.
669 * @return The ModelAction at which the next execution should diverge.
671 ModelAction * ModelChecker::get_next_backtrack()
673 ModelAction *next = priv->next_backtrack;
674 priv->next_backtrack = NULL;
679 * Processes a read or rmw model action.
680 * @param curr is the read model action to process.
681 * @param second_part_of_rmw is boolean that is true is this is the second action of a rmw.
682 * @return True if processing this read updates the mo_graph.
684 bool ModelChecker::process_read(ModelAction *curr, bool second_part_of_rmw)
686 uint64_t value = VALUE_NONE;
687 bool updated = false;
689 const ModelAction *reads_from = curr->get_node()->get_read_from();
690 if (reads_from != NULL) {
691 mo_graph->startChanges();
693 value = reads_from->get_value();
694 bool r_status = false;
696 if (!second_part_of_rmw) {
697 check_recency(curr, reads_from);
698 r_status = r_modification_order(curr, reads_from);
702 if (!second_part_of_rmw&&is_infeasible()&&(curr->get_node()->increment_read_from()||curr->get_node()->increment_future_value())) {
703 mo_graph->rollbackChanges();
704 priv->too_many_reads = false;
708 curr->read_from(reads_from);
709 mo_graph->commitChanges();
710 mo_check_promises(curr->get_tid(), reads_from);
713 } else if (!second_part_of_rmw) {
714 /* Read from future value */
715 value = curr->get_node()->get_future_value();
716 modelclock_t expiration = curr->get_node()->get_future_value_expiration();
717 curr->read_from(NULL);
718 Promise *valuepromise = new Promise(curr, value, expiration);
719 promises->push_back(valuepromise);
721 get_thread(curr)->set_return_value(value);
727 * Processes a lock, trylock, or unlock model action. @param curr is
728 * the read model action to process.
730 * The try lock operation checks whether the lock is taken. If not,
731 * it falls to the normal lock operation case. If so, it returns
734 * The lock operation has already been checked that it is enabled, so
735 * it just grabs the lock and synchronizes with the previous unlock.
737 * The unlock operation has to re-enable all of the threads that are
738 * waiting on the lock.
740 * @return True if synchronization was updated; false otherwise
742 bool ModelChecker::process_mutex(ModelAction *curr) {
743 std::mutex *mutex=NULL;
744 struct std::mutex_state *state=NULL;
746 if (curr->is_trylock() || curr->is_lock() || curr->is_unlock()) {
747 mutex = (std::mutex *)curr->get_location();
748 state = mutex->get_state();
749 } else if(curr->is_wait()) {
750 mutex = (std::mutex *)curr->get_value();
751 state = mutex->get_state();
754 switch (curr->get_type()) {
755 case ATOMIC_TRYLOCK: {
756 bool success = !state->islocked;
757 curr->set_try_lock(success);
759 get_thread(curr)->set_return_value(0);
762 get_thread(curr)->set_return_value(1);
764 //otherwise fall into the lock case
766 if (curr->get_cv()->getClock(state->alloc_tid) <= state->alloc_clock)
767 assert_bug("Lock access before initialization");
768 state->islocked = true;
769 ModelAction *unlock = get_last_unlock(curr);
770 //synchronize with the previous unlock statement
771 if (unlock != NULL) {
772 curr->synchronize_with(unlock);
777 case ATOMIC_UNLOCK: {
779 state->islocked = false;
780 //wake up the other threads
781 action_list_t *waiters = get_safe_ptr_action(lock_waiters_map, curr->get_location());
782 //activate all the waiting threads
783 for (action_list_t::iterator rit = waiters->begin(); rit != waiters->end(); rit++) {
784 scheduler->wake(get_thread(*rit));
791 state->islocked = false;
792 //wake up the other threads
793 action_list_t *waiters = get_safe_ptr_action(lock_waiters_map, (void *) curr->get_value());
794 //activate all the waiting threads
795 for (action_list_t::iterator rit = waiters->begin(); rit != waiters->end(); rit++) {
796 scheduler->wake(get_thread(*rit));
799 //check whether we should go to sleep or not...simulate spurious failures
800 if (curr->get_node()->get_misc()==0) {
801 get_safe_ptr_action(condvar_waiters_map, curr->get_location())->push_back(curr);
803 scheduler->sleep(get_current_thread());
807 case ATOMIC_NOTIFY_ALL: {
808 action_list_t *waiters = get_safe_ptr_action(condvar_waiters_map, curr->get_location());
809 //activate all the waiting threads
810 for (action_list_t::iterator rit = waiters->begin(); rit != waiters->end(); rit++) {
811 scheduler->wake(get_thread(*rit));
816 case ATOMIC_NOTIFY_ONE: {
817 action_list_t *waiters = get_safe_ptr_action(condvar_waiters_map, curr->get_location());
818 int wakeupthread=curr->get_node()->get_misc();
819 action_list_t::iterator it = waiters->begin();
820 advance(it, wakeupthread);
821 scheduler->wake(get_thread(*it));
833 * Process a write ModelAction
834 * @param curr The ModelAction to process
835 * @return True if the mo_graph was updated or promises were resolved
837 bool ModelChecker::process_write(ModelAction *curr)
839 bool updated_mod_order = w_modification_order(curr);
840 bool updated_promises = resolve_promises(curr);
842 if (promises->size() == 0) {
843 for (unsigned int i = 0; i < futurevalues->size(); i++) {
844 struct PendingFutureValue pfv = (*futurevalues)[i];
845 //Do more ambitious checks now that mo is more complete
846 if (mo_may_allow(pfv.writer, pfv.act)&&
847 pfv.act->get_node()->add_future_value(pfv.writer->get_value(), pfv.writer->get_seq_number()+params.maxfuturedelay) &&
848 (!priv->next_backtrack || *pfv.act > *priv->next_backtrack))
849 priv->next_backtrack = pfv.act;
851 futurevalues->resize(0);
854 mo_graph->commitChanges();
855 mo_check_promises(curr->get_tid(), curr);
857 get_thread(curr)->set_return_value(VALUE_NONE);
858 return updated_mod_order || updated_promises;
862 * @brief Process the current action for thread-related activity
864 * Performs current-action processing for a THREAD_* ModelAction. Proccesses
865 * may include setting Thread status, completing THREAD_FINISH/THREAD_JOIN
866 * synchronization, etc. This function is a no-op for non-THREAD actions
867 * (e.g., ATOMIC_{READ,WRITE,RMW,LOCK}, etc.)
869 * @param curr The current action
870 * @return True if synchronization was updated or a thread completed
872 bool ModelChecker::process_thread_action(ModelAction *curr)
874 bool updated = false;
876 switch (curr->get_type()) {
877 case THREAD_CREATE: {
878 Thread *th = (Thread *)curr->get_location();
879 th->set_creation(curr);
883 Thread *blocking = (Thread *)curr->get_location();
884 ModelAction *act = get_last_action(blocking->get_id());
885 curr->synchronize_with(act);
886 updated = true; /* trigger rel-seq checks */
889 case THREAD_FINISH: {
890 Thread *th = get_thread(curr);
891 while (!th->wait_list_empty()) {
892 ModelAction *act = th->pop_wait_list();
893 scheduler->wake(get_thread(act));
896 updated = true; /* trigger rel-seq checks */
900 check_promises(curr->get_tid(), NULL, curr->get_cv());
911 * @brief Process the current action for release sequence fixup activity
913 * Performs model-checker release sequence fixups for the current action,
914 * forcing a single pending release sequence to break (with a given, potential
915 * "loose" write) or to complete (i.e., synchronize). If a pending release
916 * sequence forms a complete release sequence, then we must perform the fixup
917 * synchronization, mo_graph additions, etc.
919 * @param curr The current action; must be a release sequence fixup action
920 * @param work_queue The work queue to which to add work items as they are
923 void ModelChecker::process_relseq_fixup(ModelAction *curr, work_queue_t *work_queue)
925 const ModelAction *write = curr->get_node()->get_relseq_break();
926 struct release_seq *sequence = pending_rel_seqs->back();
927 pending_rel_seqs->pop_back();
929 ModelAction *acquire = sequence->acquire;
930 const ModelAction *rf = sequence->rf;
931 const ModelAction *release = sequence->release;
935 ASSERT(release->same_thread(rf));
939 * @todo Forcing a synchronization requires that we set
940 * modification order constraints. For instance, we can't allow
941 * a fixup sequence in which two separate read-acquire
942 * operations read from the same sequence, where the first one
943 * synchronizes and the other doesn't. Essentially, we can't
944 * allow any writes to insert themselves between 'release' and
948 /* Must synchronize */
949 if (!acquire->synchronize_with(release)) {
950 set_bad_synchronization();
953 /* Re-check all pending release sequences */
954 work_queue->push_back(CheckRelSeqWorkEntry(NULL));
955 /* Re-check act for mo_graph edges */
956 work_queue->push_back(MOEdgeWorkEntry(acquire));
958 /* propagate synchronization to later actions */
959 action_list_t::reverse_iterator rit = action_trace->rbegin();
960 for (; (*rit) != acquire; rit++) {
961 ModelAction *propagate = *rit;
962 if (acquire->happens_before(propagate)) {
963 propagate->synchronize_with(acquire);
964 /* Re-check 'propagate' for mo_graph edges */
965 work_queue->push_back(MOEdgeWorkEntry(propagate));
969 /* Break release sequence with new edges:
970 * release --mo--> write --mo--> rf */
971 mo_graph->addEdge(release, write);
972 mo_graph->addEdge(write, rf);
975 /* See if we have realized a data race */
980 * Initialize the current action by performing one or more of the following
981 * actions, as appropriate: merging RMWR and RMWC/RMW actions, stepping forward
982 * in the NodeStack, manipulating backtracking sets, allocating and
983 * initializing clock vectors, and computing the promises to fulfill.
985 * @param curr The current action, as passed from the user context; may be
986 * freed/invalidated after the execution of this function, with a different
987 * action "returned" its place (pass-by-reference)
988 * @return True if curr is a newly-explored action; false otherwise
990 bool ModelChecker::initialize_curr_action(ModelAction **curr)
992 ModelAction *newcurr;
994 if ((*curr)->is_rmwc() || (*curr)->is_rmw()) {
995 newcurr = process_rmw(*curr);
998 if (newcurr->is_rmw())
999 compute_promises(newcurr);
1005 (*curr)->set_seq_number(get_next_seq_num());
1007 newcurr = node_stack->explore_action(*curr, scheduler->get_enabled_array());
1009 /* First restore type and order in case of RMW operation */
1010 if ((*curr)->is_rmwr())
1011 newcurr->copy_typeandorder(*curr);
1013 ASSERT((*curr)->get_location() == newcurr->get_location());
1014 newcurr->copy_from_new(*curr);
1016 /* Discard duplicate ModelAction; use action from NodeStack */
1019 /* Always compute new clock vector */
1020 newcurr->create_cv(get_parent_action(newcurr->get_tid()));
1023 return false; /* Action was explored previously */
1027 /* Always compute new clock vector */
1028 newcurr->create_cv(get_parent_action(newcurr->get_tid()));
1030 /* Assign most recent release fence */
1031 newcurr->set_last_fence_release(get_last_fence_release(newcurr->get_tid()));
1034 * Perform one-time actions when pushing new ModelAction onto
1037 if (newcurr->is_write())
1038 compute_promises(newcurr);
1039 else if (newcurr->is_relseq_fixup())
1040 compute_relseq_breakwrites(newcurr);
1041 else if (newcurr->is_wait())
1042 newcurr->get_node()->set_misc_max(2);
1043 else if (newcurr->is_notify_one()) {
1044 newcurr->get_node()->set_misc_max(get_safe_ptr_action(condvar_waiters_map, newcurr->get_location())->size());
1046 return true; /* This was a new ModelAction */
1051 * @brief Check whether a model action is enabled.
1053 * Checks whether a lock or join operation would be successful (i.e., is the
1054 * lock already locked, or is the joined thread already complete). If not, put
1055 * the action in a waiter list.
1057 * @param curr is the ModelAction to check whether it is enabled.
1058 * @return a bool that indicates whether the action is enabled.
1060 bool ModelChecker::check_action_enabled(ModelAction *curr) {
1061 if (curr->is_lock()) {
1062 std::mutex * lock = (std::mutex *)curr->get_location();
1063 struct std::mutex_state * state = lock->get_state();
1064 if (state->islocked) {
1065 //Stick the action in the appropriate waiting queue
1066 get_safe_ptr_action(lock_waiters_map, curr->get_location())->push_back(curr);
1069 } else if (curr->get_type() == THREAD_JOIN) {
1070 Thread *blocking = (Thread *)curr->get_location();
1071 if (!blocking->is_complete()) {
1072 blocking->push_wait_list(curr);
1081 * Stores the ModelAction for the current thread action. Call this
1082 * immediately before switching from user- to system-context to pass
1083 * data between them.
1084 * @param act The ModelAction created by the user-thread action
1086 void ModelChecker::set_current_action(ModelAction *act) {
1087 priv->current_action = act;
1091 * This is the heart of the model checker routine. It performs model-checking
1092 * actions corresponding to a given "current action." Among other processes, it
1093 * calculates reads-from relationships, updates synchronization clock vectors,
1094 * forms a memory_order constraints graph, and handles replay/backtrack
1095 * execution when running permutations of previously-observed executions.
1097 * @param curr The current action to process
1098 * @return The next Thread that must be executed. May be NULL if ModelChecker
1099 * makes no choice (e.g., according to replay execution, combining RMW actions,
1102 Thread * ModelChecker::check_current_action(ModelAction *curr)
1105 bool second_part_of_rmw = curr->is_rmwc() || curr->is_rmw();
1107 if (!check_action_enabled(curr)) {
1108 /* Make the execution look like we chose to run this action
1109 * much later, when a lock/join can succeed */
1110 get_current_thread()->set_pending(curr);
1111 scheduler->sleep(get_current_thread());
1112 return get_next_thread(NULL);
1115 bool newly_explored = initialize_curr_action(&curr);
1117 wake_up_sleeping_actions(curr);
1119 /* Add the action to lists before any other model-checking tasks */
1120 if (!second_part_of_rmw)
1121 add_action_to_lists(curr);
1123 /* Build may_read_from set for newly-created actions */
1124 if (newly_explored && curr->is_read())
1125 build_reads_from_past(curr);
1127 /* Initialize work_queue with the "current action" work */
1128 work_queue_t work_queue(1, CheckCurrWorkEntry(curr));
1129 while (!work_queue.empty() && !has_asserted()) {
1130 WorkQueueEntry work = work_queue.front();
1131 work_queue.pop_front();
1133 switch (work.type) {
1134 case WORK_CHECK_CURR_ACTION: {
1135 ModelAction *act = work.action;
1136 bool update = false; /* update this location's release seq's */
1137 bool update_all = false; /* update all release seq's */
1139 if (process_thread_action(curr))
1142 if (act->is_read() && process_read(act, second_part_of_rmw))
1145 if (act->is_write() && process_write(act))
1148 if (act->is_mutex_op() && process_mutex(act))
1151 if (act->is_relseq_fixup())
1152 process_relseq_fixup(curr, &work_queue);
1155 work_queue.push_back(CheckRelSeqWorkEntry(NULL));
1157 work_queue.push_back(CheckRelSeqWorkEntry(act->get_location()));
1160 case WORK_CHECK_RELEASE_SEQ:
1161 resolve_release_sequences(work.location, &work_queue);
1163 case WORK_CHECK_MO_EDGES: {
1164 /** @todo Complete verification of work_queue */
1165 ModelAction *act = work.action;
1166 bool updated = false;
1168 if (act->is_read()) {
1169 const ModelAction *rf = act->get_reads_from();
1170 if (rf != NULL && r_modification_order(act, rf))
1173 if (act->is_write()) {
1174 if (w_modification_order(act))
1177 mo_graph->commitChanges();
1180 work_queue.push_back(CheckRelSeqWorkEntry(act->get_location()));
1189 check_curr_backtracking(curr);
1190 set_backtracking(curr);
1191 return get_next_thread(curr);
1194 void ModelChecker::check_curr_backtracking(ModelAction * curr) {
1195 Node *currnode = curr->get_node();
1196 Node *parnode = currnode->get_parent();
1198 if ((!parnode->backtrack_empty() ||
1199 !currnode->misc_empty() ||
1200 !currnode->read_from_empty() ||
1201 !currnode->future_value_empty() ||
1202 !currnode->promise_empty() ||
1203 !currnode->relseq_break_empty())
1204 && (!priv->next_backtrack ||
1205 *curr > *priv->next_backtrack)) {
1206 priv->next_backtrack = curr;
1210 bool ModelChecker::promises_expired() const
1212 for (unsigned int promise_index = 0; promise_index < promises->size(); promise_index++) {
1213 Promise *promise = (*promises)[promise_index];
1214 if (promise->get_expiration()<priv->used_sequence_numbers) {
1222 * This is the strongest feasibility check available.
1223 * @return whether the current trace (partial or complete) must be a prefix of
1226 bool ModelChecker::isfeasibleprefix() const
1228 return pending_rel_seqs->size() == 0 && is_feasible_prefix_ignore_relseq();
1232 * Returns whether the current completed trace is feasible, except for pending
1233 * release sequences.
1235 bool ModelChecker::is_feasible_prefix_ignore_relseq() const
1237 if (DBG_ENABLED() && promises->size() != 0)
1238 DEBUG("Infeasible: unrevolved promises\n");
1240 return !is_infeasible() && promises->size() == 0;
1244 * Check if the current partial trace is infeasible. Does not check any
1245 * end-of-execution flags, which might rule out the execution. Thus, this is
1246 * useful only for ruling an execution as infeasible.
1247 * @return whether the current partial trace is infeasible.
1249 bool ModelChecker::is_infeasible() const
1251 if (DBG_ENABLED() && mo_graph->checkForRMWViolation())
1252 DEBUG("Infeasible: RMW violation\n");
1254 return mo_graph->checkForRMWViolation() || is_infeasible_ignoreRMW();
1258 * Check If the current partial trace is infeasible, while ignoring
1259 * infeasibility related to 2 RMW's reading from the same store. It does not
1260 * check end-of-execution feasibility.
1261 * @see ModelChecker::is_infeasible
1262 * @return whether the current partial trace is infeasible, ignoring multiple
1263 * RMWs reading from the same store.
1265 bool ModelChecker::is_infeasible_ignoreRMW() const
1267 if (DBG_ENABLED()) {
1268 if (mo_graph->checkForCycles())
1269 DEBUG("Infeasible: modification order cycles\n");
1270 if (priv->failed_promise)
1271 DEBUG("Infeasible: failed promise\n");
1272 if (priv->too_many_reads)
1273 DEBUG("Infeasible: too many reads\n");
1274 if (priv->bad_synchronization)
1275 DEBUG("Infeasible: bad synchronization ordering\n");
1276 if (promises_expired())
1277 DEBUG("Infeasible: promises expired\n");
1279 return mo_graph->checkForCycles() || priv->failed_promise ||
1280 priv->too_many_reads || priv->bad_synchronization ||
1284 /** Close out a RMWR by converting previous RMWR into a RMW or READ. */
1285 ModelAction * ModelChecker::process_rmw(ModelAction *act) {
1286 ModelAction *lastread = get_last_action(act->get_tid());
1287 lastread->process_rmw(act);
1288 if (act->is_rmw() && lastread->get_reads_from()!=NULL) {
1289 mo_graph->addRMWEdge(lastread->get_reads_from(), lastread);
1290 mo_graph->commitChanges();
1296 * Checks whether a thread has read from the same write for too many times
1297 * without seeing the effects of a later write.
1300 * 1) there must a different write that we could read from that would satisfy the modification order,
1301 * 2) we must have read from the same value in excess of maxreads times, and
1302 * 3) that other write must have been in the reads_from set for maxreads times.
1304 * If so, we decide that the execution is no longer feasible.
1306 void ModelChecker::check_recency(ModelAction *curr, const ModelAction *rf) {
1307 if (params.maxreads != 0) {
1309 if (curr->get_node()->get_read_from_size() <= 1)
1311 //Must make sure that execution is currently feasible... We could
1312 //accidentally clear by rolling back
1313 if (is_infeasible())
1315 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, curr->get_location());
1316 int tid = id_to_int(curr->get_tid());
1319 if ((int)thrd_lists->size() <= tid)
1321 action_list_t *list = &(*thrd_lists)[tid];
1323 action_list_t::reverse_iterator rit = list->rbegin();
1324 /* Skip past curr */
1325 for (; (*rit) != curr; rit++)
1327 /* go past curr now */
1330 action_list_t::reverse_iterator ritcopy = rit;
1331 //See if we have enough reads from the same value
1333 for (; count < params.maxreads; rit++,count++) {
1334 if (rit==list->rend())
1336 ModelAction *act = *rit;
1337 if (!act->is_read())
1340 if (act->get_reads_from() != rf)
1342 if (act->get_node()->get_read_from_size() <= 1)
1345 for (int i = 0; i<curr->get_node()->get_read_from_size(); i++) {
1347 const ModelAction * write = curr->get_node()->get_read_from_at(i);
1349 //Need a different write
1353 /* Test to see whether this is a feasible write to read from*/
1354 mo_graph->startChanges();
1355 r_modification_order(curr, write);
1356 bool feasiblereadfrom = !is_infeasible();
1357 mo_graph->rollbackChanges();
1359 if (!feasiblereadfrom)
1363 bool feasiblewrite = true;
1364 //new we need to see if this write works for everyone
1366 for (int loop = count; loop>0; loop--,rit++) {
1367 ModelAction *act=*rit;
1368 bool foundvalue = false;
1369 for (int j = 0; j<act->get_node()->get_read_from_size(); j++) {
1370 if (act->get_node()->get_read_from_at(j)==write) {
1376 feasiblewrite = false;
1380 if (feasiblewrite) {
1381 priv->too_many_reads = true;
1389 * Updates the mo_graph with the constraints imposed from the current
1392 * Basic idea is the following: Go through each other thread and find
1393 * the lastest action that happened before our read. Two cases:
1395 * (1) The action is a write => that write must either occur before
1396 * the write we read from or be the write we read from.
1398 * (2) The action is a read => the write that that action read from
1399 * must occur before the write we read from or be the same write.
1401 * @param curr The current action. Must be a read.
1402 * @param rf The action that curr reads from. Must be a write.
1403 * @return True if modification order edges were added; false otherwise
1405 bool ModelChecker::r_modification_order(ModelAction *curr, const ModelAction *rf)
1407 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, curr->get_location());
1410 ASSERT(curr->is_read());
1412 /* Last SC fence in the current thread */
1413 ModelAction *last_sc_fence_local = get_last_seq_cst_fence(curr->get_tid(), NULL);
1415 /* Iterate over all threads */
1416 for (i = 0; i < thrd_lists->size(); i++) {
1417 /* Last SC fence in thread i */
1418 ModelAction *last_sc_fence_thread_local = NULL;
1419 if (int_to_id((int)i) != curr->get_tid())
1420 last_sc_fence_thread_local = get_last_seq_cst_fence(int_to_id(i), NULL);
1422 /* Last SC fence in thread i, before last SC fence in current thread */
1423 ModelAction *last_sc_fence_thread_before = NULL;
1424 if (last_sc_fence_local)
1425 last_sc_fence_thread_before = get_last_seq_cst_fence(int_to_id(i), last_sc_fence_local);
1427 /* Iterate over actions in thread, starting from most recent */
1428 action_list_t *list = &(*thrd_lists)[i];
1429 action_list_t::reverse_iterator rit;
1430 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1431 ModelAction *act = *rit;
1433 if (act->is_write() && act != rf && act != curr) {
1434 /* C++, Section 29.3 statement 5 */
1435 if (curr->is_seqcst() && last_sc_fence_thread_local &&
1436 *act < *last_sc_fence_thread_local) {
1437 mo_graph->addEdge(act, rf);
1440 /* C++, Section 29.3 statement 4 */
1441 else if (act->is_seqcst() && last_sc_fence_local &&
1442 *act < *last_sc_fence_local) {
1443 mo_graph->addEdge(act, rf);
1446 /* C++, Section 29.3 statement 6 */
1447 else if (last_sc_fence_thread_before &&
1448 *act < *last_sc_fence_thread_before) {
1449 mo_graph->addEdge(act, rf);
1455 * Include at most one act per-thread that "happens
1456 * before" curr. Don't consider reflexively.
1458 if (act->happens_before(curr) && act != curr) {
1459 if (act->is_write()) {
1461 mo_graph->addEdge(act, rf);
1465 const ModelAction *prevreadfrom = act->get_reads_from();
1466 //if the previous read is unresolved, keep going...
1467 if (prevreadfrom == NULL)
1470 if (rf != prevreadfrom) {
1471 mo_graph->addEdge(prevreadfrom, rf);
1483 /** This method fixes up the modification order when we resolve a
1484 * promises. The basic problem is that actions that occur after the
1485 * read curr could not property add items to the modification order
1488 * So for each thread, we find the earliest item that happens after
1489 * the read curr. This is the item we have to fix up with additional
1490 * constraints. If that action is write, we add a MO edge between
1491 * the Action rf and that action. If the action is a read, we add a
1492 * MO edge between the Action rf, and whatever the read accessed.
1494 * @param curr is the read ModelAction that we are fixing up MO edges for.
1495 * @param rf is the write ModelAction that curr reads from.
1498 void ModelChecker::post_r_modification_order(ModelAction *curr, const ModelAction *rf)
1500 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, curr->get_location());
1502 ASSERT(curr->is_read());
1504 /* Iterate over all threads */
1505 for (i = 0; i < thrd_lists->size(); i++) {
1506 /* Iterate over actions in thread, starting from most recent */
1507 action_list_t *list = &(*thrd_lists)[i];
1508 action_list_t::reverse_iterator rit;
1509 ModelAction *lastact = NULL;
1511 /* Find last action that happens after curr that is either not curr or a rmw */
1512 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1513 ModelAction *act = *rit;
1514 if (curr->happens_before(act) && (curr != act || curr->is_rmw())) {
1520 /* Include at most one act per-thread that "happens before" curr */
1521 if (lastact != NULL) {
1522 if (lastact==curr) {
1523 //Case 1: The resolved read is a RMW, and we need to make sure
1524 //that the write portion of the RMW mod order after rf
1526 mo_graph->addEdge(rf, lastact);
1527 } else if (lastact->is_read()) {
1528 //Case 2: The resolved read is a normal read and the next
1529 //operation is a read, and we need to make sure the value read
1530 //is mod ordered after rf
1532 const ModelAction *postreadfrom = lastact->get_reads_from();
1533 if (postreadfrom != NULL&&rf != postreadfrom)
1534 mo_graph->addEdge(rf, postreadfrom);
1536 //Case 3: The resolved read is a normal read and the next
1537 //operation is a write, and we need to make sure that the
1538 //write is mod ordered after rf
1540 mo_graph->addEdge(rf, lastact);
1548 * Updates the mo_graph with the constraints imposed from the current write.
1550 * Basic idea is the following: Go through each other thread and find
1551 * the lastest action that happened before our write. Two cases:
1553 * (1) The action is a write => that write must occur before
1556 * (2) The action is a read => the write that that action read from
1557 * must occur before the current write.
1559 * This method also handles two other issues:
1561 * (I) Sequential Consistency: Making sure that if the current write is
1562 * seq_cst, that it occurs after the previous seq_cst write.
1564 * (II) Sending the write back to non-synchronizing reads.
1566 * @param curr The current action. Must be a write.
1567 * @return True if modification order edges were added; false otherwise
1569 bool ModelChecker::w_modification_order(ModelAction *curr)
1571 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, curr->get_location());
1574 ASSERT(curr->is_write());
1576 if (curr->is_seqcst()) {
1577 /* We have to at least see the last sequentially consistent write,
1578 so we are initialized. */
1579 ModelAction *last_seq_cst = get_last_seq_cst_write(curr);
1580 if (last_seq_cst != NULL) {
1581 mo_graph->addEdge(last_seq_cst, curr);
1586 /* Last SC fence in the current thread */
1587 ModelAction *last_sc_fence_local = get_last_seq_cst_fence(curr->get_tid(), NULL);
1589 /* Iterate over all threads */
1590 for (i = 0; i < thrd_lists->size(); i++) {
1591 /* Last SC fence in thread i, before last SC fence in current thread */
1592 ModelAction *last_sc_fence_thread_before = NULL;
1593 if (last_sc_fence_local && int_to_id((int)i) != curr->get_tid())
1594 last_sc_fence_thread_before = get_last_seq_cst_fence(int_to_id(i), last_sc_fence_local);
1596 /* Iterate over actions in thread, starting from most recent */
1597 action_list_t *list = &(*thrd_lists)[i];
1598 action_list_t::reverse_iterator rit;
1599 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1600 ModelAction *act = *rit;
1603 * 1) If RMW and it actually read from something, then we
1604 * already have all relevant edges, so just skip to next
1607 * 2) If RMW and it didn't read from anything, we should
1608 * whatever edge we can get to speed up convergence.
1610 * 3) If normal write, we need to look at earlier actions, so
1611 * continue processing list.
1613 if (curr->is_rmw()) {
1614 if (curr->get_reads_from()!=NULL)
1622 /* C++, Section 29.3 statement 7 */
1623 if (last_sc_fence_thread_before && act->is_write() &&
1624 *act < *last_sc_fence_thread_before) {
1625 mo_graph->addEdge(act, curr);
1630 * Include at most one act per-thread that "happens
1633 if (act->happens_before(curr)) {
1635 * Note: if act is RMW, just add edge:
1637 * The following edge should be handled elsewhere:
1638 * readfrom(act) --mo--> act
1640 if (act->is_write())
1641 mo_graph->addEdge(act, curr);
1642 else if (act->is_read()) {
1643 //if previous read accessed a null, just keep going
1644 if (act->get_reads_from() == NULL)
1646 mo_graph->addEdge(act->get_reads_from(), curr);
1650 } else if (act->is_read() && !act->could_synchronize_with(curr) &&
1651 !act->same_thread(curr)) {
1652 /* We have an action that:
1653 (1) did not happen before us
1654 (2) is a read and we are a write
1655 (3) cannot synchronize with us
1656 (4) is in a different thread
1658 that read could potentially read from our write. Note that
1659 these checks are overly conservative at this point, we'll
1660 do more checks before actually removing the
1664 if (thin_air_constraint_may_allow(curr, act)) {
1665 if (!is_infeasible() ||
1666 (curr->is_rmw() && act->is_rmw() && curr->get_reads_from() == act->get_reads_from() && !is_infeasible_ignoreRMW())) {
1667 struct PendingFutureValue pfv = {curr,act};
1668 futurevalues->push_back(pfv);
1678 /** Arbitrary reads from the future are not allowed. Section 29.3
1679 * part 9 places some constraints. This method checks one result of constraint
1680 * constraint. Others require compiler support. */
1681 bool ModelChecker::thin_air_constraint_may_allow(const ModelAction * writer, const ModelAction *reader) {
1682 if (!writer->is_rmw())
1685 if (!reader->is_rmw())
1688 for (const ModelAction *search = writer->get_reads_from(); search != NULL; search = search->get_reads_from()) {
1689 if (search == reader)
1691 if (search->get_tid() == reader->get_tid() &&
1692 search->happens_before(reader))
1700 * Arbitrary reads from the future are not allowed. Section 29.3 part 9 places
1701 * some constraints. This method checks one the following constraint (others
1702 * require compiler support):
1704 * If X --hb-> Y --mo-> Z, then X should not read from Z.
1706 bool ModelChecker::mo_may_allow(const ModelAction *writer, const ModelAction *reader)
1708 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, reader->get_location());
1710 /* Iterate over all threads */
1711 for (i = 0; i < thrd_lists->size(); i++) {
1712 const ModelAction *write_after_read = NULL;
1714 /* Iterate over actions in thread, starting from most recent */
1715 action_list_t *list = &(*thrd_lists)[i];
1716 action_list_t::reverse_iterator rit;
1717 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1718 ModelAction *act = *rit;
1720 if (!reader->happens_before(act))
1722 else if (act->is_write())
1723 write_after_read = act;
1724 else if (act->is_read() && act->get_reads_from() != NULL && act != reader) {
1725 write_after_read = act->get_reads_from();
1729 if (write_after_read && write_after_read!=writer && mo_graph->checkReachable(write_after_read, writer))
1736 * Finds the head(s) of the release sequence(s) containing a given ModelAction.
1737 * The ModelAction under consideration is expected to be taking part in
1738 * release/acquire synchronization as an object of the "reads from" relation.
1739 * Note that this can only provide release sequence support for RMW chains
1740 * which do not read from the future, as those actions cannot be traced until
1741 * their "promise" is fulfilled. Similarly, we may not even establish the
1742 * presence of a release sequence with certainty, as some modification order
1743 * constraints may be decided further in the future. Thus, this function
1744 * "returns" two pieces of data: a pass-by-reference vector of @a release_heads
1745 * and a boolean representing certainty.
1747 * @param rf The action that might be part of a release sequence. Must be a
1749 * @param release_heads A pass-by-reference style return parameter. After
1750 * execution of this function, release_heads will contain the heads of all the
1751 * relevant release sequences, if any exists with certainty
1752 * @param pending A pass-by-reference style return parameter which is only used
1753 * when returning false (i.e., uncertain). Returns most information regarding
1754 * an uncertain release sequence, including any write operations that might
1755 * break the sequence.
1756 * @return true, if the ModelChecker is certain that release_heads is complete;
1759 bool ModelChecker::release_seq_heads(const ModelAction *rf,
1760 rel_heads_list_t *release_heads,
1761 struct release_seq *pending) const
1763 /* Only check for release sequences if there are no cycles */
1764 if (mo_graph->checkForCycles())
1768 ASSERT(rf->is_write());
1770 if (rf->is_release())
1771 release_heads->push_back(rf);
1773 break; /* End of RMW chain */
1775 /** @todo Need to be smarter here... In the linux lock
1776 * example, this will run to the beginning of the program for
1778 /** @todo The way to be smarter here is to keep going until 1
1779 * thread has a release preceded by an acquire and you've seen
1782 /* acq_rel RMW is a sufficient stopping condition */
1783 if (rf->is_acquire() && rf->is_release())
1784 return true; /* complete */
1786 rf = rf->get_reads_from();
1789 /* read from future: need to settle this later */
1791 return false; /* incomplete */
1794 if (rf->is_release())
1795 return true; /* complete */
1797 /* else relaxed write; check modification order for contiguous subsequence
1798 * -> rf must be same thread as release */
1799 int tid = id_to_int(rf->get_tid());
1800 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, rf->get_location());
1801 action_list_t *list = &(*thrd_lists)[tid];
1802 action_list_t::const_reverse_iterator rit;
1804 /* Find rf in the thread list */
1805 rit = std::find(list->rbegin(), list->rend(), rf);
1806 ASSERT(rit != list->rend());
1808 /* Find the last write/release */
1809 for (; rit != list->rend(); rit++)
1810 if ((*rit)->is_release())
1812 if (rit == list->rend()) {
1813 /* No write-release in this thread */
1814 return true; /* complete */
1816 ModelAction *release = *rit;
1818 ASSERT(rf->same_thread(release));
1820 pending->writes.clear();
1822 bool certain = true;
1823 for (unsigned int i = 0; i < thrd_lists->size(); i++) {
1824 if (id_to_int(rf->get_tid()) == (int)i)
1826 list = &(*thrd_lists)[i];
1828 /* Can we ensure no future writes from this thread may break
1829 * the release seq? */
1830 bool future_ordered = false;
1832 ModelAction *last = get_last_action(int_to_id(i));
1833 Thread *th = get_thread(int_to_id(i));
1834 if ((last && rf->happens_before(last)) ||
1837 future_ordered = true;
1839 ASSERT(!th->is_model_thread() || future_ordered);
1841 for (rit = list->rbegin(); rit != list->rend(); rit++) {
1842 const ModelAction *act = *rit;
1843 /* Reach synchronization -> this thread is complete */
1844 if (act->happens_before(release))
1846 if (rf->happens_before(act)) {
1847 future_ordered = true;
1851 /* Only non-RMW writes can break release sequences */
1852 if (!act->is_write() || act->is_rmw())
1855 /* Check modification order */
1856 if (mo_graph->checkReachable(rf, act)) {
1857 /* rf --mo--> act */
1858 future_ordered = true;
1861 if (mo_graph->checkReachable(act, release))
1862 /* act --mo--> release */
1864 if (mo_graph->checkReachable(release, act) &&
1865 mo_graph->checkReachable(act, rf)) {
1866 /* release --mo-> act --mo--> rf */
1867 return true; /* complete */
1869 /* act may break release sequence */
1870 pending->writes.push_back(act);
1873 if (!future_ordered)
1874 certain = false; /* This thread is uncertain */
1878 release_heads->push_back(release);
1879 pending->writes.clear();
1881 pending->release = release;
1888 * A public interface for getting the release sequence head(s) with which a
1889 * given ModelAction must synchronize. This function only returns a non-empty
1890 * result when it can locate a release sequence head with certainty. Otherwise,
1891 * it may mark the internal state of the ModelChecker so that it will handle
1892 * the release sequence at a later time, causing @a act to update its
1893 * synchronization at some later point in execution.
1894 * @param act The 'acquire' action that may read from a release sequence
1895 * @param release_heads A pass-by-reference return parameter. Will be filled
1896 * with the head(s) of the release sequence(s), if they exists with certainty.
1897 * @see ModelChecker::release_seq_heads
1899 void ModelChecker::get_release_seq_heads(ModelAction *act, rel_heads_list_t *release_heads)
1901 const ModelAction *rf = act->get_reads_from();
1902 struct release_seq *sequence = (struct release_seq *)snapshot_calloc(1, sizeof(struct release_seq));
1903 sequence->acquire = act;
1905 if (!release_seq_heads(rf, release_heads, sequence)) {
1906 /* add act to 'lazy checking' list */
1907 pending_rel_seqs->push_back(sequence);
1909 snapshot_free(sequence);
1914 * Attempt to resolve all stashed operations that might synchronize with a
1915 * release sequence for a given location. This implements the "lazy" portion of
1916 * determining whether or not a release sequence was contiguous, since not all
1917 * modification order information is present at the time an action occurs.
1919 * @param location The location/object that should be checked for release
1920 * sequence resolutions. A NULL value means to check all locations.
1921 * @param work_queue The work queue to which to add work items as they are
1923 * @return True if any updates occurred (new synchronization, new mo_graph
1926 bool ModelChecker::resolve_release_sequences(void *location, work_queue_t *work_queue)
1928 bool updated = false;
1929 std::vector< struct release_seq *, SnapshotAlloc<struct release_seq *> >::iterator it = pending_rel_seqs->begin();
1930 while (it != pending_rel_seqs->end()) {
1931 struct release_seq *pending = *it;
1932 ModelAction *act = pending->acquire;
1934 /* Only resolve sequences on the given location, if provided */
1935 if (location && act->get_location() != location) {
1940 const ModelAction *rf = act->get_reads_from();
1941 rel_heads_list_t release_heads;
1943 complete = release_seq_heads(rf, &release_heads, pending);
1944 for (unsigned int i = 0; i < release_heads.size(); i++) {
1945 if (!act->has_synchronized_with(release_heads[i])) {
1946 if (act->synchronize_with(release_heads[i]))
1949 set_bad_synchronization();
1954 /* Re-check all pending release sequences */
1955 work_queue->push_back(CheckRelSeqWorkEntry(NULL));
1956 /* Re-check act for mo_graph edges */
1957 work_queue->push_back(MOEdgeWorkEntry(act));
1959 /* propagate synchronization to later actions */
1960 action_list_t::reverse_iterator rit = action_trace->rbegin();
1961 for (; (*rit) != act; rit++) {
1962 ModelAction *propagate = *rit;
1963 if (act->happens_before(propagate)) {
1964 propagate->synchronize_with(act);
1965 /* Re-check 'propagate' for mo_graph edges */
1966 work_queue->push_back(MOEdgeWorkEntry(propagate));
1971 it = pending_rel_seqs->erase(it);
1972 snapshot_free(pending);
1978 // If we resolved promises or data races, see if we have realized a data race.
1985 * Performs various bookkeeping operations for the current ModelAction. For
1986 * instance, adds action to the per-object, per-thread action vector and to the
1987 * action trace list of all thread actions.
1989 * @param act is the ModelAction to add.
1991 void ModelChecker::add_action_to_lists(ModelAction *act)
1993 int tid = id_to_int(act->get_tid());
1994 action_trace->push_back(act);
1996 get_safe_ptr_action(obj_map, act->get_location())->push_back(act);
1998 std::vector<action_list_t> *vec = get_safe_ptr_vect_action(obj_thrd_map, act->get_location());
1999 if (tid >= (int)vec->size())
2000 vec->resize(priv->next_thread_id);
2001 (*vec)[tid].push_back(act);
2003 if ((int)thrd_last_action->size() <= tid)
2004 thrd_last_action->resize(get_num_threads());
2005 (*thrd_last_action)[tid] = act;
2007 if (act->is_fence() && act->is_release()) {
2008 if ((int)thrd_last_fence_release->size() <= tid)
2009 thrd_last_fence_release->resize(get_num_threads());
2010 (*thrd_last_fence_release)[tid] = act;
2013 if (act->is_wait()) {
2014 void *mutex_loc=(void *) act->get_value();
2015 get_safe_ptr_action(obj_map, mutex_loc)->push_back(act);
2017 std::vector<action_list_t> *vec = get_safe_ptr_vect_action(obj_thrd_map, mutex_loc);
2018 if (tid >= (int)vec->size())
2019 vec->resize(priv->next_thread_id);
2020 (*vec)[tid].push_back(act);
2025 * @brief Get the last action performed by a particular Thread
2026 * @param tid The thread ID of the Thread in question
2027 * @return The last action in the thread
2029 ModelAction * ModelChecker::get_last_action(thread_id_t tid) const
2031 int threadid = id_to_int(tid);
2032 if (threadid < (int)thrd_last_action->size())
2033 return (*thrd_last_action)[id_to_int(tid)];
2039 * @brief Get the last fence release performed by a particular Thread
2040 * @param tid The thread ID of the Thread in question
2041 * @return The last fence release in the thread, if one exists; NULL otherwise
2043 ModelAction * ModelChecker::get_last_fence_release(thread_id_t tid) const
2045 int threadid = id_to_int(tid);
2046 if (threadid < (int)thrd_last_fence_release->size())
2047 return (*thrd_last_fence_release)[id_to_int(tid)];
2053 * Gets the last memory_order_seq_cst write (in the total global sequence)
2054 * performed on a particular object (i.e., memory location), not including the
2056 * @param curr The current ModelAction; also denotes the object location to
2058 * @return The last seq_cst write
2060 ModelAction * ModelChecker::get_last_seq_cst_write(ModelAction *curr) const
2062 void *location = curr->get_location();
2063 action_list_t *list = get_safe_ptr_action(obj_map, location);
2064 /* Find: max({i in dom(S) | seq_cst(t_i) && isWrite(t_i) && samevar(t_i, t)}) */
2065 action_list_t::reverse_iterator rit;
2066 for (rit = list->rbegin(); rit != list->rend(); rit++)
2067 if ((*rit)->is_write() && (*rit)->is_seqcst() && (*rit) != curr)
2073 * Gets the last memory_order_seq_cst fence (in the total global sequence)
2074 * performed in a particular thread, prior to a particular fence.
2075 * @param tid The ID of the thread to check
2076 * @param before_fence The fence from which to begin the search; if NULL, then
2077 * search for the most recent fence in the thread.
2078 * @return The last prior seq_cst fence in the thread, if exists; otherwise, NULL
2080 ModelAction * ModelChecker::get_last_seq_cst_fence(thread_id_t tid, const ModelAction *before_fence) const
2082 /* All fences should have NULL location */
2083 action_list_t *list = get_safe_ptr_action(obj_map, NULL);
2084 action_list_t::reverse_iterator rit = list->rbegin();
2087 for (; rit != list->rend(); rit++)
2088 if (*rit == before_fence)
2091 ASSERT(*rit == before_fence);
2095 for (; rit != list->rend(); rit++)
2096 if ((*rit)->is_fence() && (tid == (*rit)->get_tid()) && (*rit)->is_seqcst())
2102 * Gets the last unlock operation performed on a particular mutex (i.e., memory
2103 * location). This function identifies the mutex according to the current
2104 * action, which is presumed to perform on the same mutex.
2105 * @param curr The current ModelAction; also denotes the object location to
2107 * @return The last unlock operation
2109 ModelAction * ModelChecker::get_last_unlock(ModelAction *curr) const
2111 void *location = curr->get_location();
2112 action_list_t *list = get_safe_ptr_action(obj_map, location);
2113 /* Find: max({i in dom(S) | isUnlock(t_i) && samevar(t_i, t)}) */
2114 action_list_t::reverse_iterator rit;
2115 for (rit = list->rbegin(); rit != list->rend(); rit++)
2116 if ((*rit)->is_unlock() || (*rit)->is_wait())
2121 ModelAction * ModelChecker::get_parent_action(thread_id_t tid) const
2123 ModelAction *parent = get_last_action(tid);
2125 parent = get_thread(tid)->get_creation();
2130 * Returns the clock vector for a given thread.
2131 * @param tid The thread whose clock vector we want
2132 * @return Desired clock vector
2134 ClockVector * ModelChecker::get_cv(thread_id_t tid) const
2136 return get_parent_action(tid)->get_cv();
2140 * Resolve a set of Promises with a current write. The set is provided in the
2141 * Node corresponding to @a write.
2142 * @param write The ModelAction that is fulfilling Promises
2143 * @return True if promises were resolved; false otherwise
2145 bool ModelChecker::resolve_promises(ModelAction *write)
2147 bool resolved = false;
2148 std::vector< thread_id_t, ModelAlloc<thread_id_t> > threads_to_check;
2150 for (unsigned int i = 0, promise_index = 0; promise_index < promises->size(); i++) {
2151 Promise *promise = (*promises)[promise_index];
2152 if (write->get_node()->get_promise(i)) {
2153 ModelAction *read = promise->get_action();
2154 if (read->is_rmw()) {
2155 mo_graph->addRMWEdge(write, read);
2157 read->read_from(write);
2158 //First fix up the modification order for actions that happened
2160 r_modification_order(read, write);
2161 //Next fix up the modification order for actions that happened
2163 post_r_modification_order(read, write);
2164 //Make sure the promise's value matches the write's value
2165 ASSERT(promise->get_value() == write->get_value());
2168 promises->erase(promises->begin() + promise_index);
2169 threads_to_check.push_back(read->get_tid());
2176 //Check whether reading these writes has made threads unable to
2179 for(unsigned int i=0;i<threads_to_check.size();i++)
2180 mo_check_promises(threads_to_check[i], write);
2186 * Compute the set of promises that could potentially be satisfied by this
2187 * action. Note that the set computation actually appears in the Node, not in
2189 * @param curr The ModelAction that may satisfy promises
2191 void ModelChecker::compute_promises(ModelAction *curr)
2193 for (unsigned int i = 0; i < promises->size(); i++) {
2194 Promise *promise = (*promises)[i];
2195 const ModelAction *act = promise->get_action();
2196 if (!act->happens_before(curr) &&
2198 !act->could_synchronize_with(curr) &&
2199 !act->same_thread(curr) &&
2200 act->get_location() == curr->get_location() &&
2201 promise->get_value() == curr->get_value()) {
2202 curr->get_node()->set_promise(i, act->is_rmw());
2207 /** Checks promises in response to change in ClockVector Threads. */
2208 void ModelChecker::check_promises(thread_id_t tid, ClockVector *old_cv, ClockVector *merge_cv)
2210 for (unsigned int i = 0; i < promises->size(); i++) {
2211 Promise *promise = (*promises)[i];
2212 const ModelAction *act = promise->get_action();
2213 if ((old_cv == NULL || !old_cv->synchronized_since(act)) &&
2214 merge_cv->synchronized_since(act)) {
2215 if (promise->increment_threads(tid)) {
2216 //Promise has failed
2217 priv->failed_promise = true;
2224 void ModelChecker::check_promises_thread_disabled() {
2225 for (unsigned int i = 0; i < promises->size(); i++) {
2226 Promise *promise = (*promises)[i];
2227 if (promise->check_promise()) {
2228 priv->failed_promise = true;
2234 /** Checks promises in response to addition to modification order for threads.
2236 * pthread is the thread that performed the read that created the promise
2238 * pread is the read that created the promise
2240 * pwrite is either the first write to same location as pread by
2241 * pthread that is sequenced after pread or the value read by the
2242 * first read to the same lcoation as pread by pthread that is
2243 * sequenced after pread..
2245 * 1. If tid=pthread, then we check what other threads are reachable
2246 * through the mode order starting with pwrite. Those threads cannot
2247 * perform a write that will resolve the promise due to modification
2248 * order constraints.
2250 * 2. If the tid is not pthread, we check whether pwrite can reach the
2251 * action write through the modification order. If so, that thread
2252 * cannot perform a future write that will resolve the promise due to
2253 * modificatin order constraints.
2255 * @parem tid The thread that either read from the model action
2256 * write, or actually did the model action write.
2258 * @parem write The ModelAction representing the relevant write.
2261 void ModelChecker::mo_check_promises(thread_id_t tid, const ModelAction *write) {
2262 void * location = write->get_location();
2263 for (unsigned int i = 0; i < promises->size(); i++) {
2264 Promise *promise = (*promises)[i];
2265 const ModelAction *act = promise->get_action();
2267 //Is this promise on the same location?
2268 if ( act->get_location() != location )
2271 //same thread as the promise
2272 if ( act->get_tid()==tid ) {
2274 //do we have a pwrite for the promise, if not, set it
2275 if (promise->get_write() == NULL ) {
2276 promise->set_write(write);
2277 //The pwrite cannot happen before the promise
2278 if (write->happens_before(act) && (write != act)) {
2279 priv->failed_promise = true;
2283 if (mo_graph->checkPromise(write, promise)) {
2284 priv->failed_promise = true;
2289 //Don't do any lookups twice for the same thread
2290 if (promise->has_sync_thread(tid))
2293 if (promise->get_write()&&mo_graph->checkReachable(promise->get_write(), write)) {
2294 if (promise->increment_threads(tid)) {
2295 priv->failed_promise = true;
2303 * Compute the set of writes that may break the current pending release
2304 * sequence. This information is extracted from previou release sequence
2307 * @param curr The current ModelAction. Must be a release sequence fixup
2310 void ModelChecker::compute_relseq_breakwrites(ModelAction *curr)
2312 if (pending_rel_seqs->empty())
2315 struct release_seq *pending = pending_rel_seqs->back();
2316 for (unsigned int i = 0; i < pending->writes.size(); i++) {
2317 const ModelAction *write = pending->writes[i];
2318 curr->get_node()->add_relseq_break(write);
2321 /* NULL means don't break the sequence; just synchronize */
2322 curr->get_node()->add_relseq_break(NULL);
2326 * Build up an initial set of all past writes that this 'read' action may read
2327 * from. This set is determined by the clock vector's "happens before"
2329 * @param curr is the current ModelAction that we are exploring; it must be a
2332 void ModelChecker::build_reads_from_past(ModelAction *curr)
2334 std::vector<action_list_t> *thrd_lists = get_safe_ptr_vect_action(obj_thrd_map, curr->get_location());
2336 ASSERT(curr->is_read());
2338 ModelAction *last_sc_write = NULL;
2340 /* Track whether this object has been initialized */
2341 bool initialized = false;
2343 if (curr->is_seqcst()) {
2344 last_sc_write = get_last_seq_cst_write(curr);
2345 /* We have to at least see the last sequentially consistent write,
2346 so we are initialized. */
2347 if (last_sc_write != NULL)
2351 /* Iterate over all threads */
2352 for (i = 0; i < thrd_lists->size(); i++) {
2353 /* Iterate over actions in thread, starting from most recent */
2354 action_list_t *list = &(*thrd_lists)[i];
2355 action_list_t::reverse_iterator rit;
2356 for (rit = list->rbegin(); rit != list->rend(); rit++) {
2357 ModelAction *act = *rit;
2359 /* Only consider 'write' actions */
2360 if (!act->is_write() || act == curr)
2363 /* Don't consider more than one seq_cst write if we are a seq_cst read. */
2364 bool allow_read = true;
2366 if (curr->is_seqcst() && (act->is_seqcst() || (last_sc_write != NULL && act->happens_before(last_sc_write))) && act != last_sc_write)
2368 else if (curr->get_sleep_flag() && !curr->is_seqcst() && !sleep_can_read_from(curr, act))
2372 DEBUG("Adding action to may_read_from:\n");
2373 if (DBG_ENABLED()) {
2377 curr->get_node()->add_read_from(act);
2380 /* Include at most one act per-thread that "happens before" curr */
2381 if (act->happens_before(curr)) {
2389 assert_bug("May read from uninitialized atomic");
2391 if (DBG_ENABLED() || !initialized) {
2392 model_print("Reached read action:\n");
2394 model_print("Printing may_read_from\n");
2395 curr->get_node()->print_may_read_from();
2396 model_print("End printing may_read_from\n");
2400 bool ModelChecker::sleep_can_read_from(ModelAction * curr, const ModelAction *write) {
2402 Node *prevnode=write->get_node()->get_parent();
2404 bool thread_sleep=prevnode->enabled_status(curr->get_tid())==THREAD_SLEEP_SET;
2405 if (write->is_release()&&thread_sleep)
2407 if (!write->is_rmw()) {
2410 if (write->get_reads_from()==NULL)
2412 write=write->get_reads_from();
2416 static void print_list(action_list_t *list, int exec_num = -1)
2418 action_list_t::iterator it;
2420 model_print("---------------------------------------------------------------------\n");
2422 model_print("Execution %d:\n", exec_num);
2424 unsigned int hash=0;
2426 for (it = list->begin(); it != list->end(); it++) {
2428 hash=hash^(hash<<3)^((*it)->hash());
2430 model_print("HASH %u\n", hash);
2431 model_print("---------------------------------------------------------------------\n");
2434 #if SUPPORT_MOD_ORDER_DUMP
2435 void ModelChecker::dumpGraph(char *filename) {
2437 sprintf(buffer, "%s.dot",filename);
2438 FILE *file=fopen(buffer, "w");
2439 fprintf(file, "digraph %s {\n",filename);
2440 mo_graph->dumpNodes(file);
2441 ModelAction ** thread_array=(ModelAction **)model_calloc(1, sizeof(ModelAction *)*get_num_threads());
2443 for (action_list_t::iterator it = action_trace->begin(); it != action_trace->end(); it++) {
2444 ModelAction *action=*it;
2445 if (action->is_read()) {
2446 fprintf(file, "N%u [label=\"%u, T%u\"];\n", action->get_seq_number(),action->get_seq_number(), action->get_tid());
2447 if (action->get_reads_from()!=NULL)
2448 fprintf(file, "N%u -> N%u[label=\"rf\", color=red];\n", action->get_seq_number(), action->get_reads_from()->get_seq_number());
2450 if (thread_array[action->get_tid()] != NULL) {
2451 fprintf(file, "N%u -> N%u[label=\"sb\", color=blue];\n", thread_array[action->get_tid()]->get_seq_number(), action->get_seq_number());
2454 thread_array[action->get_tid()]=action;
2456 fprintf(file,"}\n");
2457 model_free(thread_array);
2462 /** @brief Prints an execution trace summary. */
2463 void ModelChecker::print_summary() const
2465 #if SUPPORT_MOD_ORDER_DUMP
2467 char buffername[100];
2468 sprintf(buffername, "exec%04u", stats.num_total);
2469 mo_graph->dumpGraphToFile(buffername);
2470 sprintf(buffername, "graph%04u", stats.num_total);
2471 dumpGraph(buffername);
2474 if (!isfeasibleprefix())
2475 model_print("INFEASIBLE EXECUTION!\n");
2476 print_list(action_trace, stats.num_total);
2481 * Add a Thread to the system for the first time. Should only be called once
2483 * @param t The Thread to add
2485 void ModelChecker::add_thread(Thread *t)
2487 thread_map->put(id_to_int(t->get_id()), t);
2488 scheduler->add_thread(t);
2492 * Removes a thread from the scheduler.
2493 * @param the thread to remove.
2495 void ModelChecker::remove_thread(Thread *t)
2497 scheduler->remove_thread(t);
2501 * @brief Get a Thread reference by its ID
2502 * @param tid The Thread's ID
2503 * @return A Thread reference
2505 Thread * ModelChecker::get_thread(thread_id_t tid) const
2507 return thread_map->get(id_to_int(tid));
2511 * @brief Get a reference to the Thread in which a ModelAction was executed
2512 * @param act The ModelAction
2513 * @return A Thread reference
2515 Thread * ModelChecker::get_thread(ModelAction *act) const
2517 return get_thread(act->get_tid());
2521 * @brief Check if a Thread is currently enabled
2522 * @param t The Thread to check
2523 * @return True if the Thread is currently enabled
2525 bool ModelChecker::is_enabled(Thread *t) const
2527 return scheduler->is_enabled(t);
2531 * @brief Check if a Thread is currently enabled
2532 * @param tid The ID of the Thread to check
2533 * @return True if the Thread is currently enabled
2535 bool ModelChecker::is_enabled(thread_id_t tid) const
2537 return scheduler->is_enabled(tid);
2541 * Switch from a user-context to the "master thread" context (a.k.a. system
2542 * context). This switch is made with the intention of exploring a particular
2543 * model-checking action (described by a ModelAction object). Must be called
2544 * from a user-thread context.
2546 * @param act The current action that will be explored. May be NULL only if
2547 * trace is exiting via an assertion (see ModelChecker::set_assert and
2548 * ModelChecker::has_asserted).
2549 * @return Return status from the 'swap' call (i.e., success/fail, 0/-1)
2551 int ModelChecker::switch_to_master(ModelAction *act)
2554 Thread *old = thread_current();
2555 set_current_action(act);
2556 old->set_state(THREAD_READY);
2557 return Thread::swap(old, &system_context);
2561 * Takes the next step in the execution, if possible.
2562 * @return Returns true (success) if a step was taken and false otherwise.
2564 bool ModelChecker::take_step() {
2568 Thread *curr = priv->current_action ? get_thread(priv->current_action) : NULL;
2570 if (curr->get_state() == THREAD_READY) {
2571 ASSERT(priv->current_action);
2573 priv->nextThread = check_current_action(priv->current_action);
2574 priv->current_action = NULL;
2576 if (curr->is_blocked() || curr->is_complete())
2577 scheduler->remove_thread(curr);
2582 Thread *next = scheduler->next_thread(priv->nextThread);
2584 /* Infeasible -> don't take any more steps */
2585 if (is_infeasible())
2587 else if (isfeasibleprefix() && have_bug_reports()) {
2592 if (params.bound != 0) {
2593 if (priv->used_sequence_numbers > params.bound) {
2598 DEBUG("(%d, %d)\n", curr ? id_to_int(curr->get_id()) : -1,
2599 next ? id_to_int(next->get_id()) : -1);
2602 * Launch end-of-execution release sequence fixups only when there are:
2604 * (1) no more user threads to run (or when execution replay chooses
2605 * the 'model_thread')
2606 * (2) pending release sequences
2607 * (3) pending assertions (i.e., data races)
2608 * (4) no pending promises
2610 if (!pending_rel_seqs->empty() && (!next || next->is_model_thread()) &&
2611 is_feasible_prefix_ignore_relseq() && !unrealizedraces.empty()) {
2612 model_print("*** WARNING: release sequence fixup action (%zu pending release seuqences) ***\n",
2613 pending_rel_seqs->size());
2614 ModelAction *fixup = new ModelAction(MODEL_FIXUP_RELSEQ,
2615 std::memory_order_seq_cst, NULL, VALUE_NONE,
2617 set_current_action(fixup);
2621 /* next == NULL -> don't take any more steps */
2625 next->set_state(THREAD_RUNNING);
2627 if (next->get_pending() != NULL) {
2628 /* restart a pending action */
2629 set_current_action(next->get_pending());
2630 next->set_pending(NULL);
2631 next->set_state(THREAD_READY);
2635 /* Return false only if swap fails with an error */
2636 return (Thread::swap(&system_context, next) == 0);
2639 /** Wrapper to run the user's main function, with appropriate arguments */
2640 void user_main_wrapper(void *)
2642 user_main(model->params.argc, model->params.argv);
2645 /** @brief Run ModelChecker for the user program */
2646 void ModelChecker::run()
2651 /* Start user program */
2652 add_thread(new Thread(&user_thread, &user_main_wrapper, NULL));
2654 /* Wait for all threads to complete */
2655 while (take_step());
2656 } while (next_execution());