1 /* ---------------------------------------------------------------------------
3 * (c) The GHC Team, 1998-2004
7 * Different GHC ways use this scheduler quite differently (see comments below)
8 * Here is the global picture:
10 * WAY Name CPP flag What's it for
11 * --------------------------------------
12 * mp GUM PARALLEL_HASKELL Parallel execution on a distrib. memory machine
13 * s SMP SMP Parallel execution on a shared memory machine
14 * mg GranSim GRAN Simulation of parallel execution
15 * md GUM/GdH DIST Distributed execution (based on GUM)
17 * --------------------------------------------------------------------------*/
20 * Version with support for distributed memory parallelism aka GUM (WAY=mp):
22 The main scheduling loop in GUM iterates until a finish message is received.
23 In that case a global flag @receivedFinish@ is set and this instance of
24 the RTS shuts down. See ghc/rts/parallel/HLComms.c:processMessages()
25 for the handling of incoming messages, such as PP_FINISH.
26 Note that in the parallel case we have a system manager that coordinates
27 different PEs, each of which are running one instance of the RTS.
28 See ghc/rts/parallel/SysMan.c for the main routine of the parallel program.
29 From this routine processes executing ghc/rts/Main.c are spawned. -- HWL
31 * Version with support for simulating parallel execution aka GranSim (WAY=mg):
33 The main scheduling code in GranSim is quite different from that in std
34 (concurrent) Haskell: while concurrent Haskell just iterates over the
35 threads in the runnable queue, GranSim is event driven, i.e. it iterates
36 over the events in the global event queue. -- HWL
39 #include "PosixSource.h"
44 #include "BlockAlloc.h"
45 #include "OSThreads.h"
49 #define COMPILING_SCHEDULER
51 #include "StgMiscClosures.h"
52 #include "Interpreter.h"
53 #include "Exception.h"
61 #include "ThreadLabels.h"
62 #include "LdvProfile.h"
65 #include "Proftimer.h"
68 #if defined(GRAN) || defined(PARALLEL_HASKELL)
69 # include "GranSimRts.h"
71 # include "ParallelRts.h"
72 # include "Parallel.h"
73 # include "ParallelDebug.h"
78 #include "Capability.h"
81 #ifdef HAVE_SYS_TYPES_H
82 #include <sys/types.h>
96 // Turn off inlining when debugging - it obfuscates things
99 # define STATIC_INLINE static
103 #define USED_IN_THREADED_RTS
105 #define USED_IN_THREADED_RTS STG_UNUSED
108 #ifdef RTS_SUPPORTS_THREADS
109 #define USED_WHEN_RTS_SUPPORTS_THREADS
111 #define USED_WHEN_RTS_SUPPORTS_THREADS STG_UNUSED
114 /* Main thread queue.
115 * Locks required: sched_mutex.
117 StgMainThread *main_threads = NULL;
121 StgTSO* ActiveTSO = NULL; /* for assigning system costs; GranSim-Light only */
122 /* rtsTime TimeOfNextEvent, EndOfTimeSlice; now in GranSim.c */
125 In GranSim we have a runnable and a blocked queue for each processor.
126 In order to minimise code changes new arrays run_queue_hds/tls
127 are created. run_queue_hd is then a short cut (macro) for
128 run_queue_hds[CurrentProc] (see GranSim.h).
131 StgTSO *run_queue_hds[MAX_PROC], *run_queue_tls[MAX_PROC];
132 StgTSO *blocked_queue_hds[MAX_PROC], *blocked_queue_tls[MAX_PROC];
133 StgTSO *ccalling_threadss[MAX_PROC];
134 /* We use the same global list of threads (all_threads) in GranSim as in
135 the std RTS (i.e. we are cheating). However, we don't use this list in
136 the GranSim specific code at the moment (so we are only potentially
142 * Locks required: sched_mutex.
144 StgTSO *run_queue_hd = NULL;
145 StgTSO *run_queue_tl = NULL;
146 StgTSO *blocked_queue_hd = NULL;
147 StgTSO *blocked_queue_tl = NULL;
148 StgTSO *blackhole_queue = NULL;
149 StgTSO *sleeping_queue = NULL; /* perhaps replace with a hash table? */
153 /* The blackhole_queue should be checked for threads to wake up. See
154 * Schedule.h for more thorough comment.
156 rtsBool blackholes_need_checking = rtsFalse;
158 /* Linked list of all threads.
159 * Used for detecting garbage collected threads.
161 StgTSO *all_threads = NULL;
163 /* When a thread performs a safe C call (_ccall_GC, using old
164 * terminology), it gets put on the suspended_ccalling_threads
165 * list. Used by the garbage collector.
167 static StgTSO *suspended_ccalling_threads;
169 /* KH: The following two flags are shared memory locations. There is no need
170 to lock them, since they are only unset at the end of a scheduler
174 /* flag set by signal handler to precipitate a context switch */
175 int context_switch = 0;
177 /* if this flag is set as well, give up execution */
178 rtsBool interrupted = rtsFalse;
180 /* If this flag is set, we are running Haskell code. Used to detect
181 * uses of 'foreign import unsafe' that should be 'safe'.
183 static rtsBool in_haskell = rtsFalse;
185 /* Next thread ID to allocate.
186 * Locks required: thread_id_mutex
188 static StgThreadID next_thread_id = 1;
191 * Pointers to the state of the current thread.
192 * Rule of thumb: if CurrentTSO != NULL, then we're running a Haskell
193 * thread. If CurrentTSO == NULL, then we're at the scheduler level.
196 /* The smallest stack size that makes any sense is:
197 * RESERVED_STACK_WORDS (so we can get back from the stack overflow)
198 * + sizeofW(StgStopFrame) (the stg_stop_thread_info frame)
199 * + 1 (the closure to enter)
201 * + 1 (spare slot req'd by stg_ap_v_ret)
203 * A thread with this stack will bomb immediately with a stack
204 * overflow, which will increase its stack size.
207 #define MIN_STACK_WORDS (RESERVED_STACK_WORDS + sizeofW(StgStopFrame) + 3)
214 /* This is used in `TSO.h' and gcc 2.96 insists that this variable actually
215 * exists - earlier gccs apparently didn't.
221 static Condition gc_pending_cond = INIT_COND_VAR;
224 static rtsBool ready_to_gc;
227 * Set to TRUE when entering a shutdown state (via shutdownHaskellAndExit()) --
228 * in an MT setting, needed to signal that a worker thread shouldn't hang around
229 * in the scheduler when it is out of work.
231 static rtsBool shutting_down_scheduler = rtsFalse;
233 #if defined(RTS_SUPPORTS_THREADS)
234 /* ToDo: carefully document the invariants that go together
235 * with these synchronisation objects.
237 Mutex sched_mutex = INIT_MUTEX_VAR;
238 Mutex term_mutex = INIT_MUTEX_VAR;
240 #endif /* RTS_SUPPORTS_THREADS */
242 #if defined(PARALLEL_HASKELL)
244 rtsTime TimeOfLastYield;
245 rtsBool emitSchedule = rtsTrue;
249 static char *whatNext_strs[] = {
259 /* -----------------------------------------------------------------------------
260 * static function prototypes
261 * -------------------------------------------------------------------------- */
263 #if defined(RTS_SUPPORTS_THREADS)
264 static void taskStart(void);
267 static void schedule( StgMainThread *mainThread USED_WHEN_RTS_SUPPORTS_THREADS,
268 Capability *initialCapability );
271 // These function all encapsulate parts of the scheduler loop, and are
272 // abstracted only to make the structure and control flow of the
273 // scheduler clearer.
275 static void schedulePreLoop(void);
276 static void scheduleHandleInterrupt(void);
277 static void scheduleStartSignalHandlers(void);
278 static void scheduleCheckBlockedThreads(void);
279 static void scheduleCheckBlackHoles(void);
280 static void scheduleDetectDeadlock(void);
282 static StgTSO *scheduleProcessEvent(rtsEvent *event);
284 #if defined(PARALLEL_HASKELL)
285 static StgTSO *scheduleSendPendingMessages(void);
286 static void scheduleActivateSpark(void);
287 static rtsBool scheduleGetRemoteWork(rtsBool *receivedFinish);
289 #if defined(PAR) || defined(GRAN)
290 static void scheduleGranParReport(void);
292 static void schedulePostRunThread(void);
293 static rtsBool scheduleHandleHeapOverflow( Capability *cap, StgTSO *t );
294 static void scheduleHandleStackOverflow( StgTSO *t);
295 static rtsBool scheduleHandleYield( StgTSO *t, nat prev_what_next );
296 static void scheduleHandleThreadBlocked( StgTSO *t );
297 static rtsBool scheduleHandleThreadFinished( StgMainThread *mainThread,
298 Capability *cap, StgTSO *t );
299 static void scheduleDoHeapProfile(void);
300 static void scheduleDoGC(void);
302 static void unblockThread(StgTSO *tso);
303 static rtsBool checkBlackHoles(void);
304 static SchedulerStatus waitThread_(/*out*/StgMainThread* m,
305 Capability *initialCapability
307 static void scheduleThread_ (StgTSO* tso);
308 static void AllRoots(evac_fn evac);
310 static StgTSO *threadStackOverflow(StgTSO *tso);
312 static void raiseAsync_(StgTSO *tso, StgClosure *exception,
313 rtsBool stop_at_atomically);
315 static void printThreadBlockage(StgTSO *tso);
316 static void printThreadStatus(StgTSO *tso);
318 #if defined(PARALLEL_HASKELL)
319 StgTSO * createSparkThread(rtsSpark spark);
320 StgTSO * activateSpark (rtsSpark spark);
323 /* ----------------------------------------------------------------------------
325 * ------------------------------------------------------------------------- */
327 #if defined(RTS_SUPPORTS_THREADS)
328 static rtsBool startingWorkerThread = rtsFalse;
333 ACQUIRE_LOCK(&sched_mutex);
334 startingWorkerThread = rtsFalse;
336 RELEASE_LOCK(&sched_mutex);
340 startSchedulerTaskIfNecessary(void)
342 if(run_queue_hd != END_TSO_QUEUE
343 || blocked_queue_hd != END_TSO_QUEUE
344 || sleeping_queue != END_TSO_QUEUE)
346 if(!startingWorkerThread)
347 { // we don't want to start another worker thread
348 // just because the last one hasn't yet reached the
349 // "waiting for capability" state
350 startingWorkerThread = rtsTrue;
351 if (!startTask(taskStart)) {
352 startingWorkerThread = rtsFalse;
359 /* -----------------------------------------------------------------------------
360 * Putting a thread on the run queue: different scheduling policies
361 * -------------------------------------------------------------------------- */
364 addToRunQueue( StgTSO *t )
366 #if defined(PARALLEL_HASKELL)
367 if (RtsFlags.ParFlags.doFairScheduling) {
368 // this does round-robin scheduling; good for concurrency
369 APPEND_TO_RUN_QUEUE(t);
371 // this does unfair scheduling; good for parallelism
372 PUSH_ON_RUN_QUEUE(t);
375 // this does round-robin scheduling; good for concurrency
376 APPEND_TO_RUN_QUEUE(t);
380 /* ---------------------------------------------------------------------------
381 Main scheduling loop.
383 We use round-robin scheduling, each thread returning to the
384 scheduler loop when one of these conditions is detected:
387 * timer expires (thread yields)
392 Locking notes: we acquire the scheduler lock once at the beginning
393 of the scheduler loop, and release it when
395 * running a thread, or
396 * waiting for work, or
397 * waiting for a GC to complete.
400 In a GranSim setup this loop iterates over the global event queue.
401 This revolves around the global event queue, which determines what
402 to do next. Therefore, it's more complicated than either the
403 concurrent or the parallel (GUM) setup.
406 GUM iterates over incoming messages.
407 It starts with nothing to do (thus CurrentTSO == END_TSO_QUEUE),
408 and sends out a fish whenever it has nothing to do; in-between
409 doing the actual reductions (shared code below) it processes the
410 incoming messages and deals with delayed operations
411 (see PendingFetches).
412 This is not the ugliest code you could imagine, but it's bloody close.
414 ------------------------------------------------------------------------ */
417 schedule( StgMainThread *mainThread USED_WHEN_RTS_SUPPORTS_THREADS,
418 Capability *initialCapability )
422 StgThreadReturnCode ret;
425 #elif defined(PARALLEL_HASKELL)
428 rtsBool receivedFinish = rtsFalse;
430 nat tp_size, sp_size; // stats only
435 // Pre-condition: sched_mutex is held.
436 // We might have a capability, passed in as initialCapability.
437 cap = initialCapability;
439 #if !defined(RTS_SUPPORTS_THREADS)
440 // simply initialise it in the non-threaded case
441 grabCapability(&cap);
445 sched_belch("### NEW SCHEDULER LOOP (main thr: %p, cap: %p)",
446 mainThread, initialCapability);
451 // -----------------------------------------------------------
452 // Scheduler loop starts here:
454 #if defined(PARALLEL_HASKELL)
455 #define TERMINATION_CONDITION (!receivedFinish)
457 #define TERMINATION_CONDITION ((event = get_next_event()) != (rtsEvent*)NULL)
459 #define TERMINATION_CONDITION rtsTrue
462 while (TERMINATION_CONDITION) {
465 /* Choose the processor with the next event */
466 CurrentProc = event->proc;
467 CurrentTSO = event->tso;
470 IF_DEBUG(scheduler, printAllThreads());
474 // Wait until GC has completed, if necessary.
478 releaseCapability(cap);
479 IF_DEBUG(scheduler,sched_belch("waiting for GC"));
480 waitCondition( &gc_pending_cond, &sched_mutex );
485 #if defined(RTS_SUPPORTS_THREADS)
486 // Yield the capability to higher-priority tasks if necessary.
489 yieldCapability(&cap);
492 // If we do not currently hold a capability, we wait for one
495 waitForCapability(&sched_mutex, &cap,
496 mainThread ? &mainThread->bound_thread_cond : NULL);
499 // We now have a capability...
502 // Check whether we have re-entered the RTS from Haskell without
503 // going via suspendThread()/resumeThread (i.e. a 'safe' foreign
506 errorBelch("schedule: re-entered unsafely.\n"
507 " Perhaps a 'foreign import unsafe' should be 'safe'?");
511 scheduleHandleInterrupt();
513 #if defined(not_yet) && defined(SMP)
515 // Top up the run queue from our spark pool. We try to make the
516 // number of threads in the run queue equal to the number of
517 // free capabilities.
521 if (EMPTY_RUN_QUEUE()) {
522 spark = findSpark(rtsFalse);
524 break; /* no more sparks in the pool */
526 createSparkThread(spark);
528 sched_belch("==^^ turning spark of closure %p into a thread",
529 (StgClosure *)spark));
535 scheduleStartSignalHandlers();
537 // Only check the black holes here if we've nothing else to do.
538 // During normal execution, the black hole list only gets checked
539 // at GC time, to avoid repeatedly traversing this possibly long
540 // list each time around the scheduler.
541 if (EMPTY_RUN_QUEUE()) { scheduleCheckBlackHoles(); }
543 scheduleCheckBlockedThreads();
545 scheduleDetectDeadlock();
547 // Normally, the only way we can get here with no threads to
548 // run is if a keyboard interrupt received during
549 // scheduleCheckBlockedThreads() or scheduleDetectDeadlock().
550 // Additionally, it is not fatal for the
551 // threaded RTS to reach here with no threads to run.
553 // win32: might be here due to awaitEvent() being abandoned
554 // as a result of a console event having been delivered.
555 if ( EMPTY_RUN_QUEUE() ) {
556 #if !defined(RTS_SUPPORTS_THREADS) && !defined(mingw32_HOST_OS)
559 continue; // nothing to do
562 #if defined(PARALLEL_HASKELL)
563 scheduleSendPendingMessages();
564 if (EMPTY_RUN_QUEUE() && scheduleActivateSpark())
568 ASSERT(next_fish_to_send_at==0); // i.e. no delayed fishes left!
571 /* If we still have no work we need to send a FISH to get a spark
573 if (EMPTY_RUN_QUEUE()) {
574 if (!scheduleGetRemoteWork(&receivedFinish)) continue;
575 ASSERT(rtsFalse); // should not happen at the moment
577 // from here: non-empty run queue.
578 // TODO: merge above case with this, only one call processMessages() !
579 if (PacketsWaiting()) { /* process incoming messages, if
580 any pending... only in else
581 because getRemoteWork waits for
583 receivedFinish = processMessages();
588 scheduleProcessEvent(event);
592 // Get a thread to run
594 ASSERT(run_queue_hd != END_TSO_QUEUE);
597 #if defined(GRAN) || defined(PAR)
598 scheduleGranParReport(); // some kind of debuging output
600 // Sanity check the thread we're about to run. This can be
601 // expensive if there is lots of thread switching going on...
602 IF_DEBUG(sanity,checkTSO(t));
605 #if defined(RTS_SUPPORTS_THREADS)
606 // Check whether we can run this thread in the current task.
607 // If not, we have to pass our capability to the right task.
609 StgMainThread *m = t->main;
616 sched_belch("### Running thread %d in bound thread", t->id));
617 // yes, the Haskell thread is bound to the current native thread
622 sched_belch("### thread %d bound to another OS thread", t->id));
623 // no, bound to a different Haskell thread: pass to that thread
624 PUSH_ON_RUN_QUEUE(t);
625 passCapability(&m->bound_thread_cond);
631 if(mainThread != NULL)
632 // The thread we want to run is bound.
635 sched_belch("### this OS thread cannot run thread %d", t->id));
636 // no, the current native thread is bound to a different
637 // Haskell thread, so pass it to any worker thread
638 PUSH_ON_RUN_QUEUE(t);
639 passCapabilityToWorker();
646 cap->r.rCurrentTSO = t;
648 /* context switches are now initiated by the timer signal, unless
649 * the user specified "context switch as often as possible", with
652 if ((RtsFlags.ConcFlags.ctxtSwitchTicks == 0
653 && (run_queue_hd != END_TSO_QUEUE
654 || blocked_queue_hd != END_TSO_QUEUE
655 || sleeping_queue != END_TSO_QUEUE)))
660 RELEASE_LOCK(&sched_mutex);
662 IF_DEBUG(scheduler, sched_belch("-->> running thread %ld %s ...",
663 (long)t->id, whatNext_strs[t->what_next]));
665 #if defined(PROFILING)
666 startHeapProfTimer();
669 // ----------------------------------------------------------------------
670 // Run the current thread
672 prev_what_next = t->what_next;
674 errno = t->saved_errno;
675 in_haskell = rtsTrue;
677 switch (prev_what_next) {
681 /* Thread already finished, return to scheduler. */
682 ret = ThreadFinished;
686 ret = StgRun((StgFunPtr) stg_returnToStackTop, &cap->r);
689 case ThreadInterpret:
690 ret = interpretBCO(cap);
694 barf("schedule: invalid what_next field");
697 // We have run some Haskell code: there might be blackhole-blocked
698 // threads to wake up now.
699 if ( blackhole_queue != END_TSO_QUEUE ) {
700 blackholes_need_checking = rtsTrue;
703 in_haskell = rtsFalse;
705 // The TSO might have moved, eg. if it re-entered the RTS and a GC
706 // happened. So find the new location:
707 t = cap->r.rCurrentTSO;
709 // And save the current errno in this thread.
710 t->saved_errno = errno;
712 // ----------------------------------------------------------------------
714 /* Costs for the scheduler are assigned to CCS_SYSTEM */
715 #if defined(PROFILING)
720 ACQUIRE_LOCK(&sched_mutex);
722 #if defined(RTS_SUPPORTS_THREADS)
723 IF_DEBUG(scheduler,debugBelch("sched (task %p): ", osThreadId()););
724 #elif !defined(GRAN) && !defined(PARALLEL_HASKELL)
725 IF_DEBUG(scheduler,debugBelch("sched: "););
728 schedulePostRunThread();
732 ready_to_gc = scheduleHandleHeapOverflow(cap,t);
736 scheduleHandleStackOverflow(t);
740 if (scheduleHandleYield(t, prev_what_next)) {
741 // shortcut for switching between compiler/interpreter:
747 scheduleHandleThreadBlocked(t);
752 if (scheduleHandleThreadFinished(mainThread, cap, t)) return;;
756 barf("schedule: invalid thread return code %d", (int)ret);
759 scheduleDoHeapProfile();
761 } /* end of while() */
763 IF_PAR_DEBUG(verbose,
764 debugBelch("== Leaving schedule() after having received Finish\n"));
767 /* ----------------------------------------------------------------------------
768 * Setting up the scheduler loop
769 * ASSUMES: sched_mutex
770 * ------------------------------------------------------------------------- */
773 schedulePreLoop(void)
776 /* set up first event to get things going */
777 /* ToDo: assign costs for system setup and init MainTSO ! */
778 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
780 CurrentTSO, (StgClosure*)NULL, (rtsSpark*)NULL);
783 debugBelch("GRAN: Init CurrentTSO (in schedule) = %p\n",
785 G_TSO(CurrentTSO, 5));
787 if (RtsFlags.GranFlags.Light) {
788 /* Save current time; GranSim Light only */
789 CurrentTSO->gran.clock = CurrentTime[CurrentProc];
794 /* ----------------------------------------------------------------------------
795 * Deal with the interrupt flag
796 * ASSUMES: sched_mutex
797 * ------------------------------------------------------------------------- */
800 void scheduleHandleInterrupt(void)
803 // Test for interruption. If interrupted==rtsTrue, then either
804 // we received a keyboard interrupt (^C), or the scheduler is
805 // trying to shut down all the tasks (shutting_down_scheduler) in
809 if (shutting_down_scheduler) {
810 IF_DEBUG(scheduler, sched_belch("shutting down"));
811 #if defined(RTS_SUPPORTS_THREADS)
815 IF_DEBUG(scheduler, sched_belch("interrupted"));
821 /* ----------------------------------------------------------------------------
822 * Start any pending signal handlers
823 * ASSUMES: sched_mutex
824 * ------------------------------------------------------------------------- */
827 scheduleStartSignalHandlers(void)
829 #if defined(RTS_USER_SIGNALS)
830 if (signals_pending()) {
831 RELEASE_LOCK(&sched_mutex); /* ToDo: kill */
832 startSignalHandlers();
833 ACQUIRE_LOCK(&sched_mutex);
838 /* ----------------------------------------------------------------------------
839 * Check for blocked threads that can be woken up.
840 * ASSUMES: sched_mutex
841 * ------------------------------------------------------------------------- */
844 scheduleCheckBlockedThreads(void)
847 // Check whether any waiting threads need to be woken up. If the
848 // run queue is empty, and there are no other tasks running, we
849 // can wait indefinitely for something to happen.
851 if ( !EMPTY_QUEUE(blocked_queue_hd) || !EMPTY_QUEUE(sleeping_queue) )
853 #if defined(RTS_SUPPORTS_THREADS)
854 // We shouldn't be here...
855 barf("schedule: awaitEvent() in threaded RTS");
857 awaitEvent( EMPTY_RUN_QUEUE() && !blackholes_need_checking );
862 /* ----------------------------------------------------------------------------
863 * Check for threads blocked on BLACKHOLEs that can be woken up
864 * ASSUMES: sched_mutex
865 * ------------------------------------------------------------------------- */
867 scheduleCheckBlackHoles( void )
869 if ( blackholes_need_checking )
872 blackholes_need_checking = rtsFalse;
876 /* ----------------------------------------------------------------------------
877 * Detect deadlock conditions and attempt to resolve them.
878 * ASSUMES: sched_mutex
879 * ------------------------------------------------------------------------- */
882 scheduleDetectDeadlock(void)
885 * Detect deadlock: when we have no threads to run, there are no
886 * threads blocked, waiting for I/O, or sleeping, and all the
887 * other tasks are waiting for work, we must have a deadlock of
890 if ( EMPTY_THREAD_QUEUES() )
892 #if !defined(PARALLEL_HASKELL) && !defined(RTS_SUPPORTS_THREADS)
893 IF_DEBUG(scheduler, sched_belch("deadlocked, forcing major GC..."));
895 // Garbage collection can release some new threads due to
896 // either (a) finalizers or (b) threads resurrected because
897 // they are unreachable and will therefore be sent an
898 // exception. Any threads thus released will be immediately
900 GarbageCollect(GetRoots,rtsTrue);
901 if ( !EMPTY_RUN_QUEUE() ) return;
903 #if defined(RTS_USER_SIGNALS)
904 /* If we have user-installed signal handlers, then wait
905 * for signals to arrive rather then bombing out with a
908 if ( anyUserHandlers() ) {
910 sched_belch("still deadlocked, waiting for signals..."));
914 if (signals_pending()) {
915 RELEASE_LOCK(&sched_mutex);
916 startSignalHandlers();
917 ACQUIRE_LOCK(&sched_mutex);
920 // either we have threads to run, or we were interrupted:
921 ASSERT(!EMPTY_RUN_QUEUE() || interrupted);
925 /* Probably a real deadlock. Send the current main thread the
926 * Deadlock exception (or in the SMP build, send *all* main
927 * threads the deadlock exception, since none of them can make
933 switch (m->tso->why_blocked) {
934 case BlockedOnBlackHole:
935 case BlockedOnException:
937 raiseAsync(m->tso, (StgClosure *)NonTermination_closure);
940 barf("deadlock: main thread blocked in a strange way");
944 #elif defined(RTS_SUPPORTS_THREADS)
945 // ToDo: add deadlock detection in threaded RTS
946 #elif defined(PARALLEL_HASKELL)
947 // ToDo: add deadlock detection in GUM (similar to SMP) -- HWL
952 /* ----------------------------------------------------------------------------
953 * Process an event (GRAN only)
954 * ------------------------------------------------------------------------- */
958 scheduleProcessEvent(rtsEvent *event)
962 if (RtsFlags.GranFlags.Light)
963 GranSimLight_enter_system(event, &ActiveTSO); // adjust ActiveTSO etc
965 /* adjust time based on time-stamp */
966 if (event->time > CurrentTime[CurrentProc] &&
967 event->evttype != ContinueThread)
968 CurrentTime[CurrentProc] = event->time;
970 /* Deal with the idle PEs (may issue FindWork or MoveSpark events) */
971 if (!RtsFlags.GranFlags.Light)
974 IF_DEBUG(gran, debugBelch("GRAN: switch by event-type\n"));
976 /* main event dispatcher in GranSim */
977 switch (event->evttype) {
978 /* Should just be continuing execution */
980 IF_DEBUG(gran, debugBelch("GRAN: doing ContinueThread\n"));
981 /* ToDo: check assertion
982 ASSERT(run_queue_hd != (StgTSO*)NULL &&
983 run_queue_hd != END_TSO_QUEUE);
985 /* Ignore ContinueThreads for fetching threads (if synchr comm) */
986 if (!RtsFlags.GranFlags.DoAsyncFetch &&
987 procStatus[CurrentProc]==Fetching) {
988 debugBelch("ghuH: Spurious ContinueThread while Fetching ignored; TSO %d (%p) [PE %d]\n",
989 CurrentTSO->id, CurrentTSO, CurrentProc);
992 /* Ignore ContinueThreads for completed threads */
993 if (CurrentTSO->what_next == ThreadComplete) {
994 debugBelch("ghuH: found a ContinueThread event for completed thread %d (%p) [PE %d] (ignoring ContinueThread)\n",
995 CurrentTSO->id, CurrentTSO, CurrentProc);
998 /* Ignore ContinueThreads for threads that are being migrated */
999 if (PROCS(CurrentTSO)==Nowhere) {
1000 debugBelch("ghuH: trying to run the migrating TSO %d (%p) [PE %d] (ignoring ContinueThread)\n",
1001 CurrentTSO->id, CurrentTSO, CurrentProc);
1004 /* The thread should be at the beginning of the run queue */
1005 if (CurrentTSO!=run_queue_hds[CurrentProc]) {
1006 debugBelch("ghuH: TSO %d (%p) [PE %d] is not at the start of the run_queue when doing a ContinueThread\n",
1007 CurrentTSO->id, CurrentTSO, CurrentProc);
1008 break; // run the thread anyway
1011 new_event(proc, proc, CurrentTime[proc],
1013 (StgTSO*)NULL, (StgClosure*)NULL, (rtsSpark*)NULL);
1015 */ /* Catches superfluous CONTINUEs -- should be unnecessary */
1016 break; // now actually run the thread; DaH Qu'vam yImuHbej
1019 do_the_fetchnode(event);
1020 goto next_thread; /* handle next event in event queue */
1023 do_the_globalblock(event);
1024 goto next_thread; /* handle next event in event queue */
1027 do_the_fetchreply(event);
1028 goto next_thread; /* handle next event in event queue */
1030 case UnblockThread: /* Move from the blocked queue to the tail of */
1031 do_the_unblock(event);
1032 goto next_thread; /* handle next event in event queue */
1034 case ResumeThread: /* Move from the blocked queue to the tail of */
1035 /* the runnable queue ( i.e. Qu' SImqa'lu') */
1036 event->tso->gran.blocktime +=
1037 CurrentTime[CurrentProc] - event->tso->gran.blockedat;
1038 do_the_startthread(event);
1039 goto next_thread; /* handle next event in event queue */
1042 do_the_startthread(event);
1043 goto next_thread; /* handle next event in event queue */
1046 do_the_movethread(event);
1047 goto next_thread; /* handle next event in event queue */
1050 do_the_movespark(event);
1051 goto next_thread; /* handle next event in event queue */
1054 do_the_findwork(event);
1055 goto next_thread; /* handle next event in event queue */
1058 barf("Illegal event type %u\n", event->evttype);
1061 /* This point was scheduler_loop in the old RTS */
1063 IF_DEBUG(gran, debugBelch("GRAN: after main switch\n"));
1065 TimeOfLastEvent = CurrentTime[CurrentProc];
1066 TimeOfNextEvent = get_time_of_next_event();
1067 IgnoreEvents=(TimeOfNextEvent==0); // HWL HACK
1068 // CurrentTSO = ThreadQueueHd;
1070 IF_DEBUG(gran, debugBelch("GRAN: time of next event is: %ld\n",
1073 if (RtsFlags.GranFlags.Light)
1074 GranSimLight_leave_system(event, &ActiveTSO);
1076 EndOfTimeSlice = CurrentTime[CurrentProc]+RtsFlags.GranFlags.time_slice;
1079 debugBelch("GRAN: end of time-slice is %#lx\n", EndOfTimeSlice));
1081 /* in a GranSim setup the TSO stays on the run queue */
1083 /* Take a thread from the run queue. */
1084 POP_RUN_QUEUE(t); // take_off_run_queue(t);
1087 debugBelch("GRAN: About to run current thread, which is\n");
1090 context_switch = 0; // turned on via GranYield, checking events and time slice
1093 DumpGranEvent(GR_SCHEDULE, t));
1095 procStatus[CurrentProc] = Busy;
1099 /* ----------------------------------------------------------------------------
1100 * Send pending messages (PARALLEL_HASKELL only)
1101 * ------------------------------------------------------------------------- */
1103 #if defined(PARALLEL_HASKELL)
1105 scheduleSendPendingMessages(void)
1111 # if defined(PAR) // global Mem.Mgmt., omit for now
1112 if (PendingFetches != END_BF_QUEUE) {
1117 if (RtsFlags.ParFlags.BufferTime) {
1118 // if we use message buffering, we must send away all message
1119 // packets which have become too old...
1125 /* ----------------------------------------------------------------------------
1126 * Activate spark threads (PARALLEL_HASKELL only)
1127 * ------------------------------------------------------------------------- */
1129 #if defined(PARALLEL_HASKELL)
1131 scheduleActivateSpark(void)
1134 ASSERT(EMPTY_RUN_QUEUE());
1135 /* We get here if the run queue is empty and want some work.
1136 We try to turn a spark into a thread, and add it to the run queue,
1137 from where it will be picked up in the next iteration of the scheduler
1141 /* :-[ no local threads => look out for local sparks */
1142 /* the spark pool for the current PE */
1143 pool = &(cap.r.rSparks); // JB: cap = (old) MainCap
1144 if (advisory_thread_count < RtsFlags.ParFlags.maxThreads &&
1145 pool->hd < pool->tl) {
1147 * ToDo: add GC code check that we really have enough heap afterwards!!
1149 * If we're here (no runnable threads) and we have pending
1150 * sparks, we must have a space problem. Get enough space
1151 * to turn one of those pending sparks into a
1155 spark = findSpark(rtsFalse); /* get a spark */
1156 if (spark != (rtsSpark) NULL) {
1157 tso = createThreadFromSpark(spark); /* turn the spark into a thread */
1158 IF_PAR_DEBUG(fish, // schedule,
1159 debugBelch("==== schedule: Created TSO %d (%p); %d threads active\n",
1160 tso->id, tso, advisory_thread_count));
1162 if (tso==END_TSO_QUEUE) { /* failed to activate spark->back to loop */
1163 IF_PAR_DEBUG(fish, // schedule,
1164 debugBelch("==^^ failed to create thread from spark @ %lx\n",
1166 return rtsFalse; /* failed to generate a thread */
1167 } /* otherwise fall through & pick-up new tso */
1169 IF_PAR_DEBUG(fish, // schedule,
1170 debugBelch("==^^ no local sparks (spark pool contains only NFs: %d)\n",
1171 spark_queue_len(pool)));
1172 return rtsFalse; /* failed to generate a thread */
1174 return rtsTrue; /* success in generating a thread */
1175 } else { /* no more threads permitted or pool empty */
1176 return rtsFalse; /* failed to generateThread */
1179 tso = NULL; // avoid compiler warning only
1180 return rtsFalse; /* dummy in non-PAR setup */
1183 #endif // PARALLEL_HASKELL
1185 /* ----------------------------------------------------------------------------
1186 * Get work from a remote node (PARALLEL_HASKELL only)
1187 * ------------------------------------------------------------------------- */
1189 #if defined(PARALLEL_HASKELL)
1191 scheduleGetRemoteWork(rtsBool *receivedFinish)
1193 ASSERT(EMPTY_RUN_QUEUE());
1195 if (RtsFlags.ParFlags.BufferTime) {
1196 IF_PAR_DEBUG(verbose,
1197 debugBelch("...send all pending data,"));
1200 for (i=1; i<=nPEs; i++)
1201 sendImmediately(i); // send all messages away immediately
1205 //++EDEN++ idle() , i.e. send all buffers, wait for work
1206 // suppress fishing in EDEN... just look for incoming messages
1207 // (blocking receive)
1208 IF_PAR_DEBUG(verbose,
1209 debugBelch("...wait for incoming messages...\n"));
1210 *receivedFinish = processMessages(); // blocking receive...
1212 // and reenter scheduling loop after having received something
1213 // (return rtsFalse below)
1215 # else /* activate SPARKS machinery */
1216 /* We get here, if we have no work, tried to activate a local spark, but still
1217 have no work. We try to get a remote spark, by sending a FISH message.
1218 Thread migration should be added here, and triggered when a sequence of
1219 fishes returns without work. */
1220 delay = (RtsFlags.ParFlags.fishDelay!=0ll ? RtsFlags.ParFlags.fishDelay : 0ll);
1222 /* =8-[ no local sparks => look for work on other PEs */
1224 * We really have absolutely no work. Send out a fish
1225 * (there may be some out there already), and wait for
1226 * something to arrive. We clearly can't run any threads
1227 * until a SCHEDULE or RESUME arrives, and so that's what
1228 * we're hoping to see. (Of course, we still have to
1229 * respond to other types of messages.)
1231 rtsTime now = msTime() /*CURRENT_TIME*/;
1232 IF_PAR_DEBUG(verbose,
1233 debugBelch("-- now=%ld\n", now));
1234 IF_PAR_DEBUG(fish, // verbose,
1235 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
1236 (last_fish_arrived_at!=0 &&
1237 last_fish_arrived_at+delay > now)) {
1238 debugBelch("--$$ <%llu> delaying FISH until %llu (last fish %llu, delay %llu)\n",
1239 now, last_fish_arrived_at+delay,
1240 last_fish_arrived_at,
1244 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
1245 advisory_thread_count < RtsFlags.ParFlags.maxThreads) { // send a FISH, but when?
1246 if (last_fish_arrived_at==0 ||
1247 (last_fish_arrived_at+delay <= now)) { // send FISH now!
1248 /* outstandingFishes is set in sendFish, processFish;
1249 avoid flooding system with fishes via delay */
1250 next_fish_to_send_at = 0;
1252 /* ToDo: this should be done in the main scheduling loop to avoid the
1253 busy wait here; not so bad if fish delay is very small */
1254 int iq = 0; // DEBUGGING -- HWL
1255 next_fish_to_send_at = last_fish_arrived_at+delay; // remember when to send
1256 /* send a fish when ready, but process messages that arrive in the meantime */
1258 if (PacketsWaiting()) {
1260 *receivedFinish = processMessages();
1263 } while (!*receivedFinish || now<next_fish_to_send_at);
1264 // JB: This means the fish could become obsolete, if we receive
1265 // work. Better check for work again?
1266 // last line: while (!receivedFinish || !haveWork || now<...)
1267 // next line: if (receivedFinish || haveWork )
1269 if (*receivedFinish) // no need to send a FISH if we are finishing anyway
1270 return rtsFalse; // NB: this will leave scheduler loop
1271 // immediately after return!
1273 IF_PAR_DEBUG(fish, // verbose,
1274 debugBelch("--$$ <%llu> sent delayed fish (%d processMessages); active/total threads=%d/%d\n",now,iq,run_queue_len(),advisory_thread_count));
1278 // JB: IMHO, this should all be hidden inside sendFish(...)
1280 sendFish(pe, thisPE, NEW_FISH_AGE, NEW_FISH_HISTORY,
1283 // Global statistics: count no. of fishes
1284 if (RtsFlags.ParFlags.ParStats.Global &&
1285 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
1286 globalParStats.tot_fish_mess++;
1290 /* delayed fishes must have been sent by now! */
1291 next_fish_to_send_at = 0;
1294 *receivedFinish = processMessages();
1295 # endif /* SPARKS */
1298 /* NB: this function always returns rtsFalse, meaning the scheduler
1299 loop continues with the next iteration;
1301 return code means success in finding work; we enter this function
1302 if there is no local work, thus have to send a fish which takes
1303 time until it arrives with work; in the meantime we should process
1304 messages in the main loop;
1307 #endif // PARALLEL_HASKELL
1309 /* ----------------------------------------------------------------------------
1310 * PAR/GRAN: Report stats & debugging info(?)
1311 * ------------------------------------------------------------------------- */
1313 #if defined(PAR) || defined(GRAN)
1315 scheduleGranParReport(void)
1317 ASSERT(run_queue_hd != END_TSO_QUEUE);
1319 /* Take a thread from the run queue, if we have work */
1320 POP_RUN_QUEUE(t); // take_off_run_queue(END_TSO_QUEUE);
1322 /* If this TSO has got its outport closed in the meantime,
1323 * it mustn't be run. Instead, we have to clean it up as if it was finished.
1324 * It has to be marked as TH_DEAD for this purpose.
1325 * If it is TH_TERM instead, it is supposed to have finished in the normal way.
1327 JB: TODO: investigate wether state change field could be nuked
1328 entirely and replaced by the normal tso state (whatnext
1329 field). All we want to do is to kill tsos from outside.
1332 /* ToDo: write something to the log-file
1333 if (RTSflags.ParFlags.granSimStats && !sameThread)
1334 DumpGranEvent(GR_SCHEDULE, RunnableThreadsHd);
1338 /* the spark pool for the current PE */
1339 pool = &(cap.r.rSparks); // cap = (old) MainCap
1342 debugBelch("--=^ %d threads, %d sparks on [%#x]\n",
1343 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
1346 debugBelch("--=^ %d threads, %d sparks on [%#x]\n",
1347 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
1349 if (RtsFlags.ParFlags.ParStats.Full &&
1350 (t->par.sparkname != (StgInt)0) && // only log spark generated threads
1351 (emitSchedule || // forced emit
1352 (t && LastTSO && t->id != LastTSO->id))) {
1354 we are running a different TSO, so write a schedule event to log file
1355 NB: If we use fair scheduling we also have to write a deschedule
1356 event for LastTSO; with unfair scheduling we know that the
1357 previous tso has blocked whenever we switch to another tso, so
1358 we don't need it in GUM for now
1360 IF_PAR_DEBUG(fish, // schedule,
1361 debugBelch("____ scheduling spark generated thread %d (%lx) (%lx) via a forced emit\n",t->id,t,t->par.sparkname));
1363 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1364 GR_SCHEDULE, t, (StgClosure *)NULL, 0, 0);
1365 emitSchedule = rtsFalse;
1370 /* ----------------------------------------------------------------------------
1371 * After running a thread...
1372 * ASSUMES: sched_mutex
1373 * ------------------------------------------------------------------------- */
1376 schedulePostRunThread(void)
1379 /* HACK 675: if the last thread didn't yield, make sure to print a
1380 SCHEDULE event to the log file when StgRunning the next thread, even
1381 if it is the same one as before */
1383 TimeOfLastYield = CURRENT_TIME;
1386 /* some statistics gathering in the parallel case */
1388 #if defined(GRAN) || defined(PAR) || defined(EDEN)
1392 IF_DEBUG(gran, DumpGranEvent(GR_DESCHEDULE, t));
1393 globalGranStats.tot_heapover++;
1395 globalParStats.tot_heapover++;
1402 DumpGranEvent(GR_DESCHEDULE, t));
1403 globalGranStats.tot_stackover++;
1406 // DumpGranEvent(GR_DESCHEDULE, t);
1407 globalParStats.tot_stackover++;
1411 case ThreadYielding:
1414 DumpGranEvent(GR_DESCHEDULE, t));
1415 globalGranStats.tot_yields++;
1418 // DumpGranEvent(GR_DESCHEDULE, t);
1419 globalParStats.tot_yields++;
1426 debugBelch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: ",
1427 t->id, t, whatNext_strs[t->what_next], t->block_info.closure,
1428 (t->block_info.closure==(StgClosure*)NULL ? 99 : where_is(t->block_info.closure)));
1429 if (t->block_info.closure!=(StgClosure*)NULL)
1430 print_bq(t->block_info.closure);
1433 // ??? needed; should emit block before
1435 DumpGranEvent(GR_DESCHEDULE, t));
1436 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1439 ASSERT(procStatus[CurrentProc]==Busy ||
1440 ((procStatus[CurrentProc]==Fetching) &&
1441 (t->block_info.closure!=(StgClosure*)NULL)));
1442 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1443 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1444 procStatus[CurrentProc]==Fetching))
1445 procStatus[CurrentProc] = Idle;
1448 //++PAR++ blockThread() writes the event (change?)
1452 case ThreadFinished:
1456 barf("parGlobalStats: unknown return code");
1462 /* -----------------------------------------------------------------------------
1463 * Handle a thread that returned to the scheduler with ThreadHeepOverflow
1464 * ASSUMES: sched_mutex
1465 * -------------------------------------------------------------------------- */
1468 scheduleHandleHeapOverflow( Capability *cap, StgTSO *t )
1470 // did the task ask for a large block?
1471 if (cap->r.rHpAlloc > BLOCK_SIZE) {
1472 // if so, get one and push it on the front of the nursery.
1476 blocks = (lnat)BLOCK_ROUND_UP(cap->r.rHpAlloc) / BLOCK_SIZE;
1479 debugBelch("--<< thread %ld (%s) stopped: requesting a large block (size %d)\n",
1480 (long)t->id, whatNext_strs[t->what_next], blocks));
1482 // don't do this if it would push us over the
1483 // alloc_blocks_lim limit; we'll GC first.
1484 if (alloc_blocks + blocks < alloc_blocks_lim) {
1486 alloc_blocks += blocks;
1487 bd = allocGroup( blocks );
1489 // link the new group into the list
1490 bd->link = cap->r.rCurrentNursery;
1491 bd->u.back = cap->r.rCurrentNursery->u.back;
1492 if (cap->r.rCurrentNursery->u.back != NULL) {
1493 cap->r.rCurrentNursery->u.back->link = bd;
1495 ASSERT(g0s0->blocks == cap->r.rCurrentNursery &&
1496 g0s0->blocks == cap->r.rNursery);
1497 cap->r.rNursery = g0s0->blocks = bd;
1499 cap->r.rCurrentNursery->u.back = bd;
1501 // initialise it as a nursery block. We initialise the
1502 // step, gen_no, and flags field of *every* sub-block in
1503 // this large block, because this is easier than making
1504 // sure that we always find the block head of a large
1505 // block whenever we call Bdescr() (eg. evacuate() and
1506 // isAlive() in the GC would both have to do this, at
1510 for (x = bd; x < bd + blocks; x++) {
1517 // don't forget to update the block count in g0s0.
1518 g0s0->n_blocks += blocks;
1519 // This assert can be a killer if the app is doing lots
1520 // of large block allocations.
1521 ASSERT(countBlocks(g0s0->blocks) == g0s0->n_blocks);
1523 // now update the nursery to point to the new block
1524 cap->r.rCurrentNursery = bd;
1526 // we might be unlucky and have another thread get on the
1527 // run queue before us and steal the large block, but in that
1528 // case the thread will just end up requesting another large
1530 PUSH_ON_RUN_QUEUE(t);
1531 return rtsFalse; /* not actually GC'ing */
1535 /* make all the running tasks block on a condition variable,
1536 * maybe set context_switch and wait till they all pile in,
1537 * then have them wait on a GC condition variable.
1540 debugBelch("--<< thread %ld (%s) stopped: HeapOverflow\n",
1541 (long)t->id, whatNext_strs[t->what_next]));
1544 ASSERT(!is_on_queue(t,CurrentProc));
1545 #elif defined(PARALLEL_HASKELL)
1546 /* Currently we emit a DESCHEDULE event before GC in GUM.
1547 ToDo: either add separate event to distinguish SYSTEM time from rest
1548 or just nuke this DESCHEDULE (and the following SCHEDULE) */
1549 if (0 && RtsFlags.ParFlags.ParStats.Full) {
1550 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1551 GR_DESCHEDULE, t, (StgClosure *)NULL, 0, 0);
1552 emitSchedule = rtsTrue;
1556 PUSH_ON_RUN_QUEUE(t);
1558 /* actual GC is done at the end of the while loop in schedule() */
1561 /* -----------------------------------------------------------------------------
1562 * Handle a thread that returned to the scheduler with ThreadStackOverflow
1563 * ASSUMES: sched_mutex
1564 * -------------------------------------------------------------------------- */
1567 scheduleHandleStackOverflow( StgTSO *t)
1569 IF_DEBUG(scheduler,debugBelch("--<< thread %ld (%s) stopped, StackOverflow\n",
1570 (long)t->id, whatNext_strs[t->what_next]));
1571 /* just adjust the stack for this thread, then pop it back
1576 /* enlarge the stack */
1577 StgTSO *new_t = threadStackOverflow(t);
1579 /* This TSO has moved, so update any pointers to it from the
1580 * main thread stack. It better not be on any other queues...
1581 * (it shouldn't be).
1583 if (t->main != NULL) {
1584 t->main->tso = new_t;
1586 PUSH_ON_RUN_QUEUE(new_t);
1590 /* -----------------------------------------------------------------------------
1591 * Handle a thread that returned to the scheduler with ThreadYielding
1592 * ASSUMES: sched_mutex
1593 * -------------------------------------------------------------------------- */
1596 scheduleHandleYield( StgTSO *t, nat prev_what_next )
1598 // Reset the context switch flag. We don't do this just before
1599 // running the thread, because that would mean we would lose ticks
1600 // during GC, which can lead to unfair scheduling (a thread hogs
1601 // the CPU because the tick always arrives during GC). This way
1602 // penalises threads that do a lot of allocation, but that seems
1603 // better than the alternative.
1606 /* put the thread back on the run queue. Then, if we're ready to
1607 * GC, check whether this is the last task to stop. If so, wake
1608 * up the GC thread. getThread will block during a GC until the
1612 if (t->what_next != prev_what_next) {
1613 debugBelch("--<< thread %ld (%s) stopped to switch evaluators\n",
1614 (long)t->id, whatNext_strs[t->what_next]);
1616 debugBelch("--<< thread %ld (%s) stopped, yielding\n",
1617 (long)t->id, whatNext_strs[t->what_next]);
1622 //debugBelch("&& Doing sanity check on yielding TSO %ld.", t->id);
1624 ASSERT(t->link == END_TSO_QUEUE);
1626 // Shortcut if we're just switching evaluators: don't bother
1627 // doing stack squeezing (which can be expensive), just run the
1629 if (t->what_next != prev_what_next) {
1636 ASSERT(!is_on_queue(t,CurrentProc));
1639 //debugBelch("&& Doing sanity check on all ThreadQueues (and their TSOs).");
1640 checkThreadQsSanity(rtsTrue));
1647 /* add a ContinueThread event to actually process the thread */
1648 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1650 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1652 debugBelch("GRAN: eventq and runnableq after adding yielded thread to queue again:\n");
1659 /* -----------------------------------------------------------------------------
1660 * Handle a thread that returned to the scheduler with ThreadBlocked
1661 * ASSUMES: sched_mutex
1662 * -------------------------------------------------------------------------- */
1665 scheduleHandleThreadBlocked( StgTSO *t
1666 #if !defined(GRAN) && !defined(DEBUG)
1673 debugBelch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: \n",
1674 t->id, t, whatNext_strs[t->what_next], t->block_info.closure, (t->block_info.closure==(StgClosure*)NULL ? 99 : where_is(t->block_info.closure)));
1675 if (t->block_info.closure!=(StgClosure*)NULL) print_bq(t->block_info.closure));
1677 // ??? needed; should emit block before
1679 DumpGranEvent(GR_DESCHEDULE, t));
1680 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1683 ASSERT(procStatus[CurrentProc]==Busy ||
1684 ((procStatus[CurrentProc]==Fetching) &&
1685 (t->block_info.closure!=(StgClosure*)NULL)));
1686 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1687 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1688 procStatus[CurrentProc]==Fetching))
1689 procStatus[CurrentProc] = Idle;
1693 debugBelch("--<< thread %ld (%p; %s) stopped, blocking on node %p with BQ: \n",
1694 t->id, t, whatNext_strs[t->what_next], t->block_info.closure));
1697 if (t->block_info.closure!=(StgClosure*)NULL)
1698 print_bq(t->block_info.closure));
1700 /* Send a fetch (if BlockedOnGA) and dump event to log file */
1703 /* whatever we schedule next, we must log that schedule */
1704 emitSchedule = rtsTrue;
1707 /* don't need to do anything. Either the thread is blocked on
1708 * I/O, in which case we'll have called addToBlockedQueue
1709 * previously, or it's blocked on an MVar or Blackhole, in which
1710 * case it'll be on the relevant queue already.
1712 ASSERT(t->why_blocked != NotBlocked);
1714 debugBelch("--<< thread %d (%s) stopped: ",
1715 t->id, whatNext_strs[t->what_next]);
1716 printThreadBlockage(t);
1719 /* Only for dumping event to log file
1720 ToDo: do I need this in GranSim, too?
1726 /* -----------------------------------------------------------------------------
1727 * Handle a thread that returned to the scheduler with ThreadFinished
1728 * ASSUMES: sched_mutex
1729 * -------------------------------------------------------------------------- */
1732 scheduleHandleThreadFinished( StgMainThread *mainThread
1733 USED_WHEN_RTS_SUPPORTS_THREADS,
1737 /* Need to check whether this was a main thread, and if so,
1738 * return with the return value.
1740 * We also end up here if the thread kills itself with an
1741 * uncaught exception, see Exception.cmm.
1743 IF_DEBUG(scheduler,debugBelch("--++ thread %d (%s) finished\n",
1744 t->id, whatNext_strs[t->what_next]));
1747 endThread(t, CurrentProc); // clean-up the thread
1748 #elif defined(PARALLEL_HASKELL)
1749 /* For now all are advisory -- HWL */
1750 //if(t->priority==AdvisoryPriority) ??
1751 advisory_thread_count--; // JB: Caution with this counter, buggy!
1754 if(t->dist.priority==RevalPriority)
1758 # if defined(EDENOLD)
1759 // the thread could still have an outport... (BUG)
1760 if (t->eden.outport != -1) {
1761 // delete the outport for the tso which has finished...
1762 IF_PAR_DEBUG(eden_ports,
1763 debugBelch("WARNING: Scheduler removes outport %d for TSO %d.\n",
1764 t->eden.outport, t->id));
1767 // thread still in the process (HEAVY BUG! since outport has just been closed...)
1768 if (t->eden.epid != -1) {
1769 IF_PAR_DEBUG(eden_ports,
1770 debugBelch("WARNING: Scheduler removes TSO %d from process %d .\n",
1771 t->id, t->eden.epid));
1772 removeTSOfromProcess(t);
1777 if (RtsFlags.ParFlags.ParStats.Full &&
1778 !RtsFlags.ParFlags.ParStats.Suppressed)
1779 DumpEndEvent(CURRENT_PROC, t, rtsFalse /* not mandatory */);
1781 // t->par only contains statistics: left out for now...
1783 debugBelch("**** end thread: ended sparked thread %d (%lx); sparkname: %lx\n",
1784 t->id,t,t->par.sparkname));
1786 #endif // PARALLEL_HASKELL
1789 // Check whether the thread that just completed was a main
1790 // thread, and if so return with the result.
1792 // There is an assumption here that all thread completion goes
1793 // through this point; we need to make sure that if a thread
1794 // ends up in the ThreadKilled state, that it stays on the run
1795 // queue so it can be dealt with here.
1798 #if defined(RTS_SUPPORTS_THREADS)
1801 mainThread->tso == t
1805 // We are a bound thread: this must be our thread that just
1807 ASSERT(mainThread->tso == t);
1809 if (t->what_next == ThreadComplete) {
1810 if (mainThread->ret) {
1811 // NOTE: return val is tso->sp[1] (see StgStartup.hc)
1812 *(mainThread->ret) = (StgClosure *)mainThread->tso->sp[1];
1814 mainThread->stat = Success;
1816 if (mainThread->ret) {
1817 *(mainThread->ret) = NULL;
1820 mainThread->stat = Interrupted;
1822 mainThread->stat = Killed;
1826 removeThreadLabel((StgWord)mainThread->tso->id);
1828 if (mainThread->prev == NULL) {
1829 main_threads = mainThread->link;
1831 mainThread->prev->link = mainThread->link;
1833 if (mainThread->link != NULL) {
1834 mainThread->link->prev = NULL;
1836 releaseCapability(cap);
1837 return rtsTrue; // tells schedule() to return
1840 #ifdef RTS_SUPPORTS_THREADS
1841 ASSERT(t->main == NULL);
1843 if (t->main != NULL) {
1844 // Must be a main thread that is not the topmost one. Leave
1845 // it on the run queue until the stack has unwound to the
1846 // point where we can deal with this. Leaving it on the run
1847 // queue also ensures that the garbage collector knows about
1848 // this thread and its return value (it gets dropped from the
1849 // all_threads list so there's no other way to find it).
1850 APPEND_TO_RUN_QUEUE(t);
1856 /* -----------------------------------------------------------------------------
1857 * Perform a heap census, if PROFILING
1858 * -------------------------------------------------------------------------- */
1861 scheduleDoHeapProfile(void)
1864 // When we have +RTS -i0 and we're heap profiling, do a census at
1865 // every GC. This lets us get repeatable runs for debugging.
1866 if (performHeapProfile ||
1867 (RtsFlags.ProfFlags.profileInterval==0 &&
1868 RtsFlags.ProfFlags.doHeapProfile && ready_to_gc)) {
1869 GarbageCollect(GetRoots, rtsTrue);
1871 performHeapProfile = rtsFalse;
1872 ready_to_gc = rtsFalse; // we already GC'd
1877 /* -----------------------------------------------------------------------------
1878 * Perform a garbage collection if necessary
1879 * ASSUMES: sched_mutex
1880 * -------------------------------------------------------------------------- */
1888 // The last task to stop actually gets to do the GC. The rest
1889 // of the tasks release their capabilities and wait gc_pending_cond.
1890 if (ready_to_gc && allFreeCapabilities())
1895 /* Kick any transactions which are invalid back to their
1896 * atomically frames. When next scheduled they will try to
1897 * commit, this commit will fail and they will retry.
1899 for (t = all_threads; t != END_TSO_QUEUE; t = t -> link) {
1900 if (t -> what_next != ThreadRelocated && t -> trec != NO_TREC && t -> why_blocked == NotBlocked) {
1901 if (!stmValidateTransaction (t -> trec)) {
1902 IF_DEBUG(stm, sched_belch("trec %p found wasting its time", t));
1904 // strip the stack back to the ATOMICALLY_FRAME, aborting
1905 // the (nested) transaction, and saving the stack of any
1906 // partially-evaluated thunks on the heap.
1907 raiseAsync_(t, NULL, rtsTrue);
1910 ASSERT(get_itbl((StgClosure *)t->sp)->type == ATOMICALLY_FRAME);
1916 // so this happens periodically:
1917 scheduleCheckBlackHoles();
1919 /* everybody back, start the GC.
1920 * Could do it in this thread, or signal a condition var
1921 * to do it in another thread. Either way, we need to
1922 * broadcast on gc_pending_cond afterward.
1924 #if defined(RTS_SUPPORTS_THREADS)
1925 IF_DEBUG(scheduler,sched_belch("doing GC"));
1927 GarbageCollect(GetRoots,rtsFalse);
1928 ready_to_gc = rtsFalse;
1930 broadcastCondition(&gc_pending_cond);
1933 /* add a ContinueThread event to continue execution of current thread */
1934 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1936 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1938 debugBelch("GRAN: eventq and runnableq after Garbage collection:\n\n");
1945 /* ---------------------------------------------------------------------------
1946 * rtsSupportsBoundThreads(): is the RTS built to support bound threads?
1947 * used by Control.Concurrent for error checking.
1948 * ------------------------------------------------------------------------- */
1951 rtsSupportsBoundThreads(void)
1960 /* ---------------------------------------------------------------------------
1961 * isThreadBound(tso): check whether tso is bound to an OS thread.
1962 * ------------------------------------------------------------------------- */
1965 isThreadBound(StgTSO* tso USED_IN_THREADED_RTS)
1968 return (tso->main != NULL);
1973 /* ---------------------------------------------------------------------------
1974 * Singleton fork(). Do not copy any running threads.
1975 * ------------------------------------------------------------------------- */
1977 #ifndef mingw32_HOST_OS
1978 #define FORKPROCESS_PRIMOP_SUPPORTED
1981 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1983 deleteThreadImmediately(StgTSO *tso);
1986 forkProcess(HsStablePtr *entry
1987 #ifndef FORKPROCESS_PRIMOP_SUPPORTED
1992 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1998 IF_DEBUG(scheduler,sched_belch("forking!"));
1999 rts_lock(); // This not only acquires sched_mutex, it also
2000 // makes sure that no other threads are running
2004 if (pid) { /* parent */
2006 /* just return the pid */
2010 } else { /* child */
2013 // delete all threads
2014 run_queue_hd = run_queue_tl = END_TSO_QUEUE;
2016 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
2019 // don't allow threads to catch the ThreadKilled exception
2020 deleteThreadImmediately(t);
2023 // wipe the main thread list
2024 while((m = main_threads) != NULL) {
2025 main_threads = m->link;
2026 # ifdef THREADED_RTS
2027 closeCondition(&m->bound_thread_cond);
2032 rc = rts_evalStableIO(entry, NULL); // run the action
2033 rts_checkSchedStatus("forkProcess",rc);
2037 hs_exit(); // clean up and exit
2040 #else /* !FORKPROCESS_PRIMOP_SUPPORTED */
2041 barf("forkProcess#: primop not supported, sorry!\n");
2046 /* ---------------------------------------------------------------------------
2047 * deleteAllThreads(): kill all the live threads.
2049 * This is used when we catch a user interrupt (^C), before performing
2050 * any necessary cleanups and running finalizers.
2052 * Locks: sched_mutex held.
2053 * ------------------------------------------------------------------------- */
2056 deleteAllThreads ( void )
2059 IF_DEBUG(scheduler,sched_belch("deleting all threads"));
2060 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
2061 next = t->global_link;
2065 // The run queue now contains a bunch of ThreadKilled threads. We
2066 // must not throw these away: the main thread(s) will be in there
2067 // somewhere, and the main scheduler loop has to deal with it.
2068 // Also, the run queue is the only thing keeping these threads from
2069 // being GC'd, and we don't want the "main thread has been GC'd" panic.
2071 ASSERT(blocked_queue_hd == END_TSO_QUEUE);
2072 ASSERT(blackhole_queue == END_TSO_QUEUE);
2073 ASSERT(sleeping_queue == END_TSO_QUEUE);
2076 /* startThread and insertThread are now in GranSim.c -- HWL */
2079 /* ---------------------------------------------------------------------------
2080 * Suspending & resuming Haskell threads.
2082 * When making a "safe" call to C (aka _ccall_GC), the task gives back
2083 * its capability before calling the C function. This allows another
2084 * task to pick up the capability and carry on running Haskell
2085 * threads. It also means that if the C call blocks, it won't lock
2088 * The Haskell thread making the C call is put to sleep for the
2089 * duration of the call, on the susepended_ccalling_threads queue. We
2090 * give out a token to the task, which it can use to resume the thread
2091 * on return from the C function.
2092 * ------------------------------------------------------------------------- */
2095 suspendThread( StgRegTable *reg )
2099 int saved_errno = errno;
2101 /* assume that *reg is a pointer to the StgRegTable part
2104 cap = (Capability *)((void *)((unsigned char*)reg - sizeof(StgFunTable)));
2106 ACQUIRE_LOCK(&sched_mutex);
2109 sched_belch("thread %d did a _ccall_gc", cap->r.rCurrentTSO->id));
2111 // XXX this might not be necessary --SDM
2112 cap->r.rCurrentTSO->what_next = ThreadRunGHC;
2114 threadPaused(cap->r.rCurrentTSO);
2115 cap->r.rCurrentTSO->link = suspended_ccalling_threads;
2116 suspended_ccalling_threads = cap->r.rCurrentTSO;
2118 if(cap->r.rCurrentTSO->blocked_exceptions == NULL) {
2119 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall;
2120 cap->r.rCurrentTSO->blocked_exceptions = END_TSO_QUEUE;
2122 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall_NoUnblockExc;
2125 /* Use the thread ID as the token; it should be unique */
2126 tok = cap->r.rCurrentTSO->id;
2128 /* Hand back capability */
2129 releaseCapability(cap);
2131 #if defined(RTS_SUPPORTS_THREADS)
2132 /* Preparing to leave the RTS, so ensure there's a native thread/task
2133 waiting to take over.
2135 IF_DEBUG(scheduler, sched_belch("worker (token %d): leaving RTS", tok));
2138 in_haskell = rtsFalse;
2139 RELEASE_LOCK(&sched_mutex);
2141 errno = saved_errno;
2146 resumeThread( StgInt tok )
2148 StgTSO *tso, **prev;
2150 int saved_errno = errno;
2152 #if defined(RTS_SUPPORTS_THREADS)
2153 /* Wait for permission to re-enter the RTS with the result. */
2154 ACQUIRE_LOCK(&sched_mutex);
2155 waitForReturnCapability(&sched_mutex, &cap);
2157 IF_DEBUG(scheduler, sched_belch("worker (token %d): re-entering RTS", tok));
2159 grabCapability(&cap);
2162 /* Remove the thread off of the suspended list */
2163 prev = &suspended_ccalling_threads;
2164 for (tso = suspended_ccalling_threads;
2165 tso != END_TSO_QUEUE;
2166 prev = &tso->link, tso = tso->link) {
2167 if (tso->id == (StgThreadID)tok) {
2172 if (tso == END_TSO_QUEUE) {
2173 barf("resumeThread: thread not found");
2175 tso->link = END_TSO_QUEUE;
2177 if(tso->why_blocked == BlockedOnCCall) {
2178 awakenBlockedQueueNoLock(tso->blocked_exceptions);
2179 tso->blocked_exceptions = NULL;
2182 /* Reset blocking status */
2183 tso->why_blocked = NotBlocked;
2185 cap->r.rCurrentTSO = tso;
2186 in_haskell = rtsTrue;
2187 RELEASE_LOCK(&sched_mutex);
2188 errno = saved_errno;
2192 /* ---------------------------------------------------------------------------
2193 * Comparing Thread ids.
2195 * This is used from STG land in the implementation of the
2196 * instances of Eq/Ord for ThreadIds.
2197 * ------------------------------------------------------------------------ */
2200 cmp_thread(StgPtr tso1, StgPtr tso2)
2202 StgThreadID id1 = ((StgTSO *)tso1)->id;
2203 StgThreadID id2 = ((StgTSO *)tso2)->id;
2205 if (id1 < id2) return (-1);
2206 if (id1 > id2) return 1;
2210 /* ---------------------------------------------------------------------------
2211 * Fetching the ThreadID from an StgTSO.
2213 * This is used in the implementation of Show for ThreadIds.
2214 * ------------------------------------------------------------------------ */
2216 rts_getThreadId(StgPtr tso)
2218 return ((StgTSO *)tso)->id;
2223 labelThread(StgPtr tso, char *label)
2228 /* Caveat: Once set, you can only set the thread name to "" */
2229 len = strlen(label)+1;
2230 buf = stgMallocBytes(len * sizeof(char), "Schedule.c:labelThread()");
2231 strncpy(buf,label,len);
2232 /* Update will free the old memory for us */
2233 updateThreadLabel(((StgTSO *)tso)->id,buf);
2237 /* ---------------------------------------------------------------------------
2238 Create a new thread.
2240 The new thread starts with the given stack size. Before the
2241 scheduler can run, however, this thread needs to have a closure
2242 (and possibly some arguments) pushed on its stack. See
2243 pushClosure() in Schedule.h.
2245 createGenThread() and createIOThread() (in SchedAPI.h) are
2246 convenient packaged versions of this function.
2248 currently pri (priority) is only used in a GRAN setup -- HWL
2249 ------------------------------------------------------------------------ */
2251 /* currently pri (priority) is only used in a GRAN setup -- HWL */
2253 createThread(nat size, StgInt pri)
2256 createThread(nat size)
2263 /* First check whether we should create a thread at all */
2264 #if defined(PARALLEL_HASKELL)
2265 /* check that no more than RtsFlags.ParFlags.maxThreads threads are created */
2266 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads) {
2268 debugBelch("{createThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)\n",
2269 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
2270 return END_TSO_QUEUE;
2276 ASSERT(!RtsFlags.GranFlags.Light || CurrentProc==0);
2279 // ToDo: check whether size = stack_size - TSO_STRUCT_SIZEW
2281 /* catch ridiculously small stack sizes */
2282 if (size < MIN_STACK_WORDS + TSO_STRUCT_SIZEW) {
2283 size = MIN_STACK_WORDS + TSO_STRUCT_SIZEW;
2286 stack_size = size - TSO_STRUCT_SIZEW;
2288 tso = (StgTSO *)allocate(size);
2289 TICK_ALLOC_TSO(stack_size, 0);
2291 SET_HDR(tso, &stg_TSO_info, CCS_SYSTEM);
2293 SET_GRAN_HDR(tso, ThisPE);
2296 // Always start with the compiled code evaluator
2297 tso->what_next = ThreadRunGHC;
2299 tso->id = next_thread_id++;
2300 tso->why_blocked = NotBlocked;
2301 tso->blocked_exceptions = NULL;
2303 tso->saved_errno = 0;
2306 tso->stack_size = stack_size;
2307 tso->max_stack_size = round_to_mblocks(RtsFlags.GcFlags.maxStkSize)
2309 tso->sp = (P_)&(tso->stack) + stack_size;
2311 tso->trec = NO_TREC;
2314 tso->prof.CCCS = CCS_MAIN;
2317 /* put a stop frame on the stack */
2318 tso->sp -= sizeofW(StgStopFrame);
2319 SET_HDR((StgClosure*)tso->sp,(StgInfoTable *)&stg_stop_thread_info,CCS_SYSTEM);
2320 tso->link = END_TSO_QUEUE;
2324 /* uses more flexible routine in GranSim */
2325 insertThread(tso, CurrentProc);
2327 /* In a non-GranSim setup the pushing of a TSO onto the runq is separated
2333 if (RtsFlags.GranFlags.GranSimStats.Full)
2334 DumpGranEvent(GR_START,tso);
2335 #elif defined(PARALLEL_HASKELL)
2336 if (RtsFlags.ParFlags.ParStats.Full)
2337 DumpGranEvent(GR_STARTQ,tso);
2338 /* HACk to avoid SCHEDULE
2342 /* Link the new thread on the global thread list.
2344 tso->global_link = all_threads;
2348 tso->dist.priority = MandatoryPriority; //by default that is...
2352 tso->gran.pri = pri;
2354 tso->gran.magic = TSO_MAGIC; // debugging only
2356 tso->gran.sparkname = 0;
2357 tso->gran.startedat = CURRENT_TIME;
2358 tso->gran.exported = 0;
2359 tso->gran.basicblocks = 0;
2360 tso->gran.allocs = 0;
2361 tso->gran.exectime = 0;
2362 tso->gran.fetchtime = 0;
2363 tso->gran.fetchcount = 0;
2364 tso->gran.blocktime = 0;
2365 tso->gran.blockcount = 0;
2366 tso->gran.blockedat = 0;
2367 tso->gran.globalsparks = 0;
2368 tso->gran.localsparks = 0;
2369 if (RtsFlags.GranFlags.Light)
2370 tso->gran.clock = Now; /* local clock */
2372 tso->gran.clock = 0;
2374 IF_DEBUG(gran,printTSO(tso));
2375 #elif defined(PARALLEL_HASKELL)
2377 tso->par.magic = TSO_MAGIC; // debugging only
2379 tso->par.sparkname = 0;
2380 tso->par.startedat = CURRENT_TIME;
2381 tso->par.exported = 0;
2382 tso->par.basicblocks = 0;
2383 tso->par.allocs = 0;
2384 tso->par.exectime = 0;
2385 tso->par.fetchtime = 0;
2386 tso->par.fetchcount = 0;
2387 tso->par.blocktime = 0;
2388 tso->par.blockcount = 0;
2389 tso->par.blockedat = 0;
2390 tso->par.globalsparks = 0;
2391 tso->par.localsparks = 0;
2395 globalGranStats.tot_threads_created++;
2396 globalGranStats.threads_created_on_PE[CurrentProc]++;
2397 globalGranStats.tot_sq_len += spark_queue_len(CurrentProc);
2398 globalGranStats.tot_sq_probes++;
2399 #elif defined(PARALLEL_HASKELL)
2400 // collect parallel global statistics (currently done together with GC stats)
2401 if (RtsFlags.ParFlags.ParStats.Global &&
2402 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
2403 //debugBelch("Creating thread %d @ %11.2f\n", tso->id, usertime());
2404 globalParStats.tot_threads_created++;
2410 sched_belch("==__ schedule: Created TSO %d (%p);",
2411 CurrentProc, tso, tso->id));
2412 #elif defined(PARALLEL_HASKELL)
2413 IF_PAR_DEBUG(verbose,
2414 sched_belch("==__ schedule: Created TSO %d (%p); %d threads active",
2415 (long)tso->id, tso, advisory_thread_count));
2417 IF_DEBUG(scheduler,sched_belch("created thread %ld, stack size = %lx words",
2418 (long)tso->id, (long)tso->stack_size));
2425 all parallel thread creation calls should fall through the following routine.
2428 createThreadFromSpark(rtsSpark spark)
2430 ASSERT(spark != (rtsSpark)NULL);
2431 // JB: TAKE CARE OF THIS COUNTER! BUGGY
2432 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads)
2434 barf("{createSparkThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
2435 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
2436 return END_TSO_QUEUE;
2440 tso = createThread(RtsFlags.GcFlags.initialStkSize);
2441 if (tso==END_TSO_QUEUE)
2442 barf("createSparkThread: Cannot create TSO");
2444 tso->priority = AdvisoryPriority;
2446 pushClosure(tso,spark);
2448 advisory_thread_count++; // JB: TAKE CARE OF THIS COUNTER! BUGGY
2455 Turn a spark into a thread.
2456 ToDo: fix for SMP (needs to acquire SCHED_MUTEX!)
2460 activateSpark (rtsSpark spark)
2464 tso = createSparkThread(spark);
2465 if (RtsFlags.ParFlags.ParStats.Full) {
2466 //ASSERT(run_queue_hd == END_TSO_QUEUE); // I think ...
2467 IF_PAR_DEBUG(verbose,
2468 debugBelch("==^^ activateSpark: turning spark of closure %p (%s) into a thread\n",
2469 (StgClosure *)spark, info_type((StgClosure *)spark)));
2471 // ToDo: fwd info on local/global spark to thread -- HWL
2472 // tso->gran.exported = spark->exported;
2473 // tso->gran.locked = !spark->global;
2474 // tso->gran.sparkname = spark->name;
2480 /* ---------------------------------------------------------------------------
2483 * scheduleThread puts a thread on the head of the runnable queue.
2484 * This will usually be done immediately after a thread is created.
2485 * The caller of scheduleThread must create the thread using e.g.
2486 * createThread and push an appropriate closure
2487 * on this thread's stack before the scheduler is invoked.
2488 * ------------------------------------------------------------------------ */
2491 scheduleThread_(StgTSO *tso)
2493 // The thread goes at the *end* of the run-queue, to avoid possible
2494 // starvation of any threads already on the queue.
2495 APPEND_TO_RUN_QUEUE(tso);
2500 scheduleThread(StgTSO* tso)
2502 ACQUIRE_LOCK(&sched_mutex);
2503 scheduleThread_(tso);
2504 RELEASE_LOCK(&sched_mutex);
2507 #if defined(RTS_SUPPORTS_THREADS)
2508 static Condition bound_cond_cache;
2509 static int bound_cond_cache_full = 0;
2514 scheduleWaitThread(StgTSO* tso, /*[out]*/HaskellObj* ret,
2515 Capability *initialCapability)
2517 // Precondition: sched_mutex must be held
2520 m = stgMallocBytes(sizeof(StgMainThread), "waitThread");
2525 m->link = main_threads;
2527 if (main_threads != NULL) {
2528 main_threads->prev = m;
2532 #if defined(RTS_SUPPORTS_THREADS)
2533 // Allocating a new condition for each thread is expensive, so we
2534 // cache one. This is a pretty feeble hack, but it helps speed up
2535 // consecutive call-ins quite a bit.
2536 if (bound_cond_cache_full) {
2537 m->bound_thread_cond = bound_cond_cache;
2538 bound_cond_cache_full = 0;
2540 initCondition(&m->bound_thread_cond);
2544 /* Put the thread on the main-threads list prior to scheduling the TSO.
2545 Failure to do so introduces a race condition in the MT case (as
2546 identified by Wolfgang Thaller), whereby the new task/OS thread
2547 created by scheduleThread_() would complete prior to the thread
2548 that spawned it managed to put 'itself' on the main-threads list.
2549 The upshot of it all being that the worker thread wouldn't get to
2550 signal the completion of the its work item for the main thread to
2551 see (==> it got stuck waiting.) -- sof 6/02.
2553 IF_DEBUG(scheduler, sched_belch("waiting for thread (%d)", tso->id));
2555 APPEND_TO_RUN_QUEUE(tso);
2556 // NB. Don't call threadRunnable() here, because the thread is
2557 // bound and only runnable by *this* OS thread, so waking up other
2558 // workers will just slow things down.
2560 return waitThread_(m, initialCapability);
2563 /* ---------------------------------------------------------------------------
2566 * Initialise the scheduler. This resets all the queues - if the
2567 * queues contained any threads, they'll be garbage collected at the
2570 * ------------------------------------------------------------------------ */
2578 for (i=0; i<=MAX_PROC; i++) {
2579 run_queue_hds[i] = END_TSO_QUEUE;
2580 run_queue_tls[i] = END_TSO_QUEUE;
2581 blocked_queue_hds[i] = END_TSO_QUEUE;
2582 blocked_queue_tls[i] = END_TSO_QUEUE;
2583 ccalling_threadss[i] = END_TSO_QUEUE;
2584 blackhole_queue[i] = END_TSO_QUEUE;
2585 sleeping_queue = END_TSO_QUEUE;
2588 run_queue_hd = END_TSO_QUEUE;
2589 run_queue_tl = END_TSO_QUEUE;
2590 blocked_queue_hd = END_TSO_QUEUE;
2591 blocked_queue_tl = END_TSO_QUEUE;
2592 blackhole_queue = END_TSO_QUEUE;
2593 sleeping_queue = END_TSO_QUEUE;
2596 suspended_ccalling_threads = END_TSO_QUEUE;
2598 main_threads = NULL;
2599 all_threads = END_TSO_QUEUE;
2604 RtsFlags.ConcFlags.ctxtSwitchTicks =
2605 RtsFlags.ConcFlags.ctxtSwitchTime / TICK_MILLISECS;
2607 #if defined(RTS_SUPPORTS_THREADS)
2608 /* Initialise the mutex and condition variables used by
2610 initMutex(&sched_mutex);
2611 initMutex(&term_mutex);
2614 ACQUIRE_LOCK(&sched_mutex);
2616 /* A capability holds the state a native thread needs in
2617 * order to execute STG code. At least one capability is
2618 * floating around (only SMP builds have more than one).
2622 #if defined(RTS_SUPPORTS_THREADS)
2623 /* start our haskell execution tasks */
2624 startTaskManager(0,taskStart);
2627 #if /* defined(SMP) ||*/ defined(PARALLEL_HASKELL)
2631 RELEASE_LOCK(&sched_mutex);
2635 exitScheduler( void )
2637 #if defined(RTS_SUPPORTS_THREADS)
2640 interrupted = rtsTrue;
2641 shutting_down_scheduler = rtsTrue;
2644 /* ----------------------------------------------------------------------------
2645 Managing the per-task allocation areas.
2647 Each capability comes with an allocation area. These are
2648 fixed-length block lists into which allocation can be done.
2650 ToDo: no support for two-space collection at the moment???
2651 ------------------------------------------------------------------------- */
2653 static SchedulerStatus
2654 waitThread_(StgMainThread* m, Capability *initialCapability)
2656 SchedulerStatus stat;
2658 // Precondition: sched_mutex must be held.
2659 IF_DEBUG(scheduler, sched_belch("new main thread (%d)", m->tso->id));
2662 /* GranSim specific init */
2663 CurrentTSO = m->tso; // the TSO to run
2664 procStatus[MainProc] = Busy; // status of main PE
2665 CurrentProc = MainProc; // PE to run it on
2666 schedule(m,initialCapability);
2668 schedule(m,initialCapability);
2669 ASSERT(m->stat != NoStatus);
2674 #if defined(RTS_SUPPORTS_THREADS)
2675 // Free the condition variable, returning it to the cache if possible.
2676 if (!bound_cond_cache_full) {
2677 bound_cond_cache = m->bound_thread_cond;
2678 bound_cond_cache_full = 1;
2680 closeCondition(&m->bound_thread_cond);
2684 IF_DEBUG(scheduler, sched_belch("main thread (%d) finished", m->tso->id));
2687 // Postcondition: sched_mutex still held
2691 /* ---------------------------------------------------------------------------
2692 Where are the roots that we know about?
2694 - all the threads on the runnable queue
2695 - all the threads on the blocked queue
2696 - all the threads on the sleeping queue
2697 - all the thread currently executing a _ccall_GC
2698 - all the "main threads"
2700 ------------------------------------------------------------------------ */
2702 /* This has to be protected either by the scheduler monitor, or by the
2703 garbage collection monitor (probably the latter).
2708 GetRoots( evac_fn evac )
2713 for (i=0; i<=RtsFlags.GranFlags.proc; i++) {
2714 if ((run_queue_hds[i] != END_TSO_QUEUE) && ((run_queue_hds[i] != NULL)))
2715 evac((StgClosure **)&run_queue_hds[i]);
2716 if ((run_queue_tls[i] != END_TSO_QUEUE) && ((run_queue_tls[i] != NULL)))
2717 evac((StgClosure **)&run_queue_tls[i]);
2719 if ((blocked_queue_hds[i] != END_TSO_QUEUE) && ((blocked_queue_hds[i] != NULL)))
2720 evac((StgClosure **)&blocked_queue_hds[i]);
2721 if ((blocked_queue_tls[i] != END_TSO_QUEUE) && ((blocked_queue_tls[i] != NULL)))
2722 evac((StgClosure **)&blocked_queue_tls[i]);
2723 if ((ccalling_threadss[i] != END_TSO_QUEUE) && ((ccalling_threadss[i] != NULL)))
2724 evac((StgClosure **)&ccalling_threads[i]);
2731 if (run_queue_hd != END_TSO_QUEUE) {
2732 ASSERT(run_queue_tl != END_TSO_QUEUE);
2733 evac((StgClosure **)&run_queue_hd);
2734 evac((StgClosure **)&run_queue_tl);
2737 if (blocked_queue_hd != END_TSO_QUEUE) {
2738 ASSERT(blocked_queue_tl != END_TSO_QUEUE);
2739 evac((StgClosure **)&blocked_queue_hd);
2740 evac((StgClosure **)&blocked_queue_tl);
2743 if (sleeping_queue != END_TSO_QUEUE) {
2744 evac((StgClosure **)&sleeping_queue);
2748 if (blackhole_queue != END_TSO_QUEUE) {
2749 evac((StgClosure **)&blackhole_queue);
2752 if (suspended_ccalling_threads != END_TSO_QUEUE) {
2753 evac((StgClosure **)&suspended_ccalling_threads);
2756 #if defined(PARALLEL_HASKELL) || defined(GRAN)
2757 markSparkQueue(evac);
2760 #if defined(RTS_USER_SIGNALS)
2761 // mark the signal handlers (signals should be already blocked)
2762 markSignalHandlers(evac);
2766 /* -----------------------------------------------------------------------------
2769 This is the interface to the garbage collector from Haskell land.
2770 We provide this so that external C code can allocate and garbage
2771 collect when called from Haskell via _ccall_GC.
2773 It might be useful to provide an interface whereby the programmer
2774 can specify more roots (ToDo).
2776 This needs to be protected by the GC condition variable above. KH.
2777 -------------------------------------------------------------------------- */
2779 static void (*extra_roots)(evac_fn);
2784 /* Obligated to hold this lock upon entry */
2785 ACQUIRE_LOCK(&sched_mutex);
2786 GarbageCollect(GetRoots,rtsFalse);
2787 RELEASE_LOCK(&sched_mutex);
2791 performMajorGC(void)
2793 ACQUIRE_LOCK(&sched_mutex);
2794 GarbageCollect(GetRoots,rtsTrue);
2795 RELEASE_LOCK(&sched_mutex);
2799 AllRoots(evac_fn evac)
2801 GetRoots(evac); // the scheduler's roots
2802 extra_roots(evac); // the user's roots
2806 performGCWithRoots(void (*get_roots)(evac_fn))
2808 ACQUIRE_LOCK(&sched_mutex);
2809 extra_roots = get_roots;
2810 GarbageCollect(AllRoots,rtsFalse);
2811 RELEASE_LOCK(&sched_mutex);
2814 /* -----------------------------------------------------------------------------
2817 If the thread has reached its maximum stack size, then raise the
2818 StackOverflow exception in the offending thread. Otherwise
2819 relocate the TSO into a larger chunk of memory and adjust its stack
2821 -------------------------------------------------------------------------- */
2824 threadStackOverflow(StgTSO *tso)
2826 nat new_stack_size, stack_words;
2831 IF_DEBUG(sanity,checkTSO(tso));
2832 if (tso->stack_size >= tso->max_stack_size) {
2835 debugBelch("@@ threadStackOverflow of TSO %ld (%p): stack too large (now %ld; max is %ld)\n",
2836 (long)tso->id, tso, (long)tso->stack_size, (long)tso->max_stack_size);
2837 /* If we're debugging, just print out the top of the stack */
2838 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2841 /* Send this thread the StackOverflow exception */
2842 raiseAsync(tso, (StgClosure *)stackOverflow_closure);
2846 /* Try to double the current stack size. If that takes us over the
2847 * maximum stack size for this thread, then use the maximum instead.
2848 * Finally round up so the TSO ends up as a whole number of blocks.
2850 new_stack_size = stg_min(tso->stack_size * 2, tso->max_stack_size);
2851 new_tso_size = (lnat)BLOCK_ROUND_UP(new_stack_size * sizeof(W_) +
2852 TSO_STRUCT_SIZE)/sizeof(W_);
2853 new_tso_size = round_to_mblocks(new_tso_size); /* Be MBLOCK-friendly */
2854 new_stack_size = new_tso_size - TSO_STRUCT_SIZEW;
2856 IF_DEBUG(scheduler, debugBelch("== sched: increasing stack size from %d words to %d.\n", tso->stack_size, new_stack_size));
2858 dest = (StgTSO *)allocate(new_tso_size);
2859 TICK_ALLOC_TSO(new_stack_size,0);
2861 /* copy the TSO block and the old stack into the new area */
2862 memcpy(dest,tso,TSO_STRUCT_SIZE);
2863 stack_words = tso->stack + tso->stack_size - tso->sp;
2864 new_sp = (P_)dest + new_tso_size - stack_words;
2865 memcpy(new_sp, tso->sp, stack_words * sizeof(W_));
2867 /* relocate the stack pointers... */
2869 dest->stack_size = new_stack_size;
2871 /* Mark the old TSO as relocated. We have to check for relocated
2872 * TSOs in the garbage collector and any primops that deal with TSOs.
2874 * It's important to set the sp value to just beyond the end
2875 * of the stack, so we don't attempt to scavenge any part of the
2878 tso->what_next = ThreadRelocated;
2880 tso->sp = (P_)&(tso->stack[tso->stack_size]);
2881 tso->why_blocked = NotBlocked;
2883 IF_PAR_DEBUG(verbose,
2884 debugBelch("@@ threadStackOverflow of TSO %d (now at %p): stack size increased to %ld\n",
2885 tso->id, tso, tso->stack_size);
2886 /* If we're debugging, just print out the top of the stack */
2887 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2890 IF_DEBUG(sanity,checkTSO(tso));
2892 IF_DEBUG(scheduler,printTSO(dest));
2898 /* ---------------------------------------------------------------------------
2899 Wake up a queue that was blocked on some resource.
2900 ------------------------------------------------------------------------ */
2904 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2907 #elif defined(PARALLEL_HASKELL)
2909 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2911 /* write RESUME events to log file and
2912 update blocked and fetch time (depending on type of the orig closure) */
2913 if (RtsFlags.ParFlags.ParStats.Full) {
2914 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
2915 GR_RESUMEQ, ((StgTSO *)bqe), ((StgTSO *)bqe)->block_info.closure,
2916 0, 0 /* spark_queue_len(ADVISORY_POOL) */);
2917 if (EMPTY_RUN_QUEUE())
2918 emitSchedule = rtsTrue;
2920 switch (get_itbl(node)->type) {
2922 ((StgTSO *)bqe)->par.fetchtime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2927 ((StgTSO *)bqe)->par.blocktime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2934 barf("{unblockOneLocked}Daq Qagh: unexpected closure in blocking queue");
2941 static StgBlockingQueueElement *
2942 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2945 PEs node_loc, tso_loc;
2947 node_loc = where_is(node); // should be lifted out of loop
2948 tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2949 tso_loc = where_is((StgClosure *)tso);
2950 if (IS_LOCAL_TO(PROCS(node),tso_loc)) { // TSO is local
2951 /* !fake_fetch => TSO is on CurrentProc is same as IS_LOCAL_TO */
2952 ASSERT(CurrentProc!=node_loc || tso_loc==CurrentProc);
2953 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.lunblocktime;
2954 // insertThread(tso, node_loc);
2955 new_event(tso_loc, tso_loc, CurrentTime[CurrentProc],
2957 tso, node, (rtsSpark*)NULL);
2958 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2961 } else { // TSO is remote (actually should be FMBQ)
2962 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.mpacktime +
2963 RtsFlags.GranFlags.Costs.gunblocktime +
2964 RtsFlags.GranFlags.Costs.latency;
2965 new_event(tso_loc, CurrentProc, CurrentTime[CurrentProc],
2967 tso, node, (rtsSpark*)NULL);
2968 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2971 /* the thread-queue-overhead is accounted for in either Resume or UnblockThread */
2973 debugBelch(" %s TSO %d (%p) [PE %d] (block_info.closure=%p) (next=%p) ,",
2974 (node_loc==tso_loc ? "Local" : "Global"),
2975 tso->id, tso, CurrentProc, tso->block_info.closure, tso->link));
2976 tso->block_info.closure = NULL;
2977 IF_DEBUG(scheduler,debugBelch("-- Waking up thread %ld (%p)\n",
2980 #elif defined(PARALLEL_HASKELL)
2981 static StgBlockingQueueElement *
2982 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2984 StgBlockingQueueElement *next;
2986 switch (get_itbl(bqe)->type) {
2988 ASSERT(((StgTSO *)bqe)->why_blocked != NotBlocked);
2989 /* if it's a TSO just push it onto the run_queue */
2991 ((StgTSO *)bqe)->link = END_TSO_QUEUE; // debugging?
2992 APPEND_TO_RUN_QUEUE((StgTSO *)bqe);
2994 unblockCount(bqe, node);
2995 /* reset blocking status after dumping event */
2996 ((StgTSO *)bqe)->why_blocked = NotBlocked;
3000 /* if it's a BLOCKED_FETCH put it on the PendingFetches list */
3002 bqe->link = (StgBlockingQueueElement *)PendingFetches;
3003 PendingFetches = (StgBlockedFetch *)bqe;
3007 /* can ignore this case in a non-debugging setup;
3008 see comments on RBHSave closures above */
3010 /* check that the closure is an RBHSave closure */
3011 ASSERT(get_itbl((StgClosure *)bqe) == &stg_RBH_Save_0_info ||
3012 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_1_info ||
3013 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_2_info);
3017 barf("{unblockOneLocked}Daq Qagh: Unexpected IP (%#lx; %s) in blocking queue at %#lx\n",
3018 get_itbl((StgClosure *)bqe), info_type((StgClosure *)bqe),
3022 IF_PAR_DEBUG(bq, debugBelch(", %p (%s)\n", bqe, info_type((StgClosure*)bqe)));
3026 #else /* !GRAN && !PARALLEL_HASKELL */
3028 unblockOneLocked(StgTSO *tso)
3032 ASSERT(get_itbl(tso)->type == TSO);
3033 ASSERT(tso->why_blocked != NotBlocked);
3034 tso->why_blocked = NotBlocked;
3036 tso->link = END_TSO_QUEUE;
3037 APPEND_TO_RUN_QUEUE(tso);
3039 IF_DEBUG(scheduler,sched_belch("waking up thread %ld", (long)tso->id));
3044 #if defined(GRAN) || defined(PARALLEL_HASKELL)
3045 INLINE_ME StgBlockingQueueElement *
3046 unblockOne(StgBlockingQueueElement *bqe, StgClosure *node)
3048 ACQUIRE_LOCK(&sched_mutex);
3049 bqe = unblockOneLocked(bqe, node);
3050 RELEASE_LOCK(&sched_mutex);
3055 unblockOne(StgTSO *tso)
3057 ACQUIRE_LOCK(&sched_mutex);
3058 tso = unblockOneLocked(tso);
3059 RELEASE_LOCK(&sched_mutex);
3066 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
3068 StgBlockingQueueElement *bqe;
3073 debugBelch("##-_ AwBQ for node %p on PE %d @ %ld by TSO %d (%p): \n", \
3074 node, CurrentProc, CurrentTime[CurrentProc],
3075 CurrentTSO->id, CurrentTSO));
3077 node_loc = where_is(node);
3079 ASSERT(q == END_BQ_QUEUE ||
3080 get_itbl(q)->type == TSO || // q is either a TSO or an RBHSave
3081 get_itbl(q)->type == CONSTR); // closure (type constructor)
3082 ASSERT(is_unique(node));
3084 /* FAKE FETCH: magically copy the node to the tso's proc;
3085 no Fetch necessary because in reality the node should not have been
3086 moved to the other PE in the first place
3088 if (CurrentProc!=node_loc) {
3090 debugBelch("## node %p is on PE %d but CurrentProc is %d (TSO %d); assuming fake fetch and adjusting bitmask (old: %#x)\n",
3091 node, node_loc, CurrentProc, CurrentTSO->id,
3092 // CurrentTSO, where_is(CurrentTSO),
3093 node->header.gran.procs));
3094 node->header.gran.procs = (node->header.gran.procs) | PE_NUMBER(CurrentProc);
3096 debugBelch("## new bitmask of node %p is %#x\n",
3097 node, node->header.gran.procs));
3098 if (RtsFlags.GranFlags.GranSimStats.Global) {
3099 globalGranStats.tot_fake_fetches++;
3104 // ToDo: check: ASSERT(CurrentProc==node_loc);
3105 while (get_itbl(bqe)->type==TSO) { // q != END_TSO_QUEUE) {
3108 bqe points to the current element in the queue
3109 next points to the next element in the queue
3111 //tso = (StgTSO *)bqe; // wastes an assignment to get the type right
3112 //tso_loc = where_is(tso);
3114 bqe = unblockOneLocked(bqe, node);
3117 /* if this is the BQ of an RBH, we have to put back the info ripped out of
3118 the closure to make room for the anchor of the BQ */
3119 if (bqe!=END_BQ_QUEUE) {
3120 ASSERT(get_itbl(node)->type == RBH && get_itbl(bqe)->type == CONSTR);
3122 ASSERT((info_ptr==&RBH_Save_0_info) ||
3123 (info_ptr==&RBH_Save_1_info) ||
3124 (info_ptr==&RBH_Save_2_info));
3126 /* cf. convertToRBH in RBH.c for writing the RBHSave closure */
3127 ((StgRBH *)node)->blocking_queue = (StgBlockingQueueElement *)((StgRBHSave *)bqe)->payload[0];
3128 ((StgRBH *)node)->mut_link = (StgMutClosure *)((StgRBHSave *)bqe)->payload[1];
3131 debugBelch("## Filled in RBH_Save for %p (%s) at end of AwBQ\n",
3132 node, info_type(node)));
3135 /* statistics gathering */
3136 if (RtsFlags.GranFlags.GranSimStats.Global) {
3137 // globalGranStats.tot_bq_processing_time += bq_processing_time;
3138 globalGranStats.tot_bq_len += len; // total length of all bqs awakened
3139 // globalGranStats.tot_bq_len_local += len_local; // same for local TSOs only
3140 globalGranStats.tot_awbq++; // total no. of bqs awakened
3143 debugBelch("## BQ Stats of %p: [%d entries] %s\n",
3144 node, len, (bqe!=END_BQ_QUEUE) ? "RBH" : ""));
3146 #elif defined(PARALLEL_HASKELL)
3148 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
3150 StgBlockingQueueElement *bqe;
3152 ACQUIRE_LOCK(&sched_mutex);
3154 IF_PAR_DEBUG(verbose,
3155 debugBelch("##-_ AwBQ for node %p on [%x]: \n",
3159 if(get_itbl(q)->type == CONSTR || q==END_BQ_QUEUE) {
3160 IF_PAR_DEBUG(verbose, debugBelch("## ... nothing to unblock so lets just return. RFP (BUG?)\n"));
3165 ASSERT(q == END_BQ_QUEUE ||
3166 get_itbl(q)->type == TSO ||
3167 get_itbl(q)->type == BLOCKED_FETCH ||
3168 get_itbl(q)->type == CONSTR);
3171 while (get_itbl(bqe)->type==TSO ||
3172 get_itbl(bqe)->type==BLOCKED_FETCH) {
3173 bqe = unblockOneLocked(bqe, node);
3175 RELEASE_LOCK(&sched_mutex);
3178 #else /* !GRAN && !PARALLEL_HASKELL */
3181 awakenBlockedQueueNoLock(StgTSO *tso)
3183 while (tso != END_TSO_QUEUE) {
3184 tso = unblockOneLocked(tso);
3189 awakenBlockedQueue(StgTSO *tso)
3191 ACQUIRE_LOCK(&sched_mutex);
3192 while (tso != END_TSO_QUEUE) {
3193 tso = unblockOneLocked(tso);
3195 RELEASE_LOCK(&sched_mutex);
3199 /* ---------------------------------------------------------------------------
3201 - usually called inside a signal handler so it mustn't do anything fancy.
3202 ------------------------------------------------------------------------ */
3205 interruptStgRts(void)
3211 /* -----------------------------------------------------------------------------
3214 This is for use when we raise an exception in another thread, which
3216 This has nothing to do with the UnblockThread event in GranSim. -- HWL
3217 -------------------------------------------------------------------------- */
3219 #if defined(GRAN) || defined(PARALLEL_HASKELL)
3221 NB: only the type of the blocking queue is different in GranSim and GUM
3222 the operations on the queue-elements are the same
3223 long live polymorphism!
3225 Locks: sched_mutex is held upon entry and exit.
3229 unblockThread(StgTSO *tso)
3231 StgBlockingQueueElement *t, **last;
3233 switch (tso->why_blocked) {
3236 return; /* not blocked */
3239 // Be careful: nothing to do here! We tell the scheduler that the thread
3240 // is runnable and we leave it to the stack-walking code to abort the
3241 // transaction while unwinding the stack. We should perhaps have a debugging
3242 // test to make sure that this really happens and that the 'zombie' transaction
3243 // does not get committed.
3247 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
3249 StgBlockingQueueElement *last_tso = END_BQ_QUEUE;
3250 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
3252 last = (StgBlockingQueueElement **)&mvar->head;
3253 for (t = (StgBlockingQueueElement *)mvar->head;
3255 last = &t->link, last_tso = t, t = t->link) {
3256 if (t == (StgBlockingQueueElement *)tso) {
3257 *last = (StgBlockingQueueElement *)tso->link;
3258 if (mvar->tail == tso) {
3259 mvar->tail = (StgTSO *)last_tso;
3264 barf("unblockThread (MVAR): TSO not found");
3267 case BlockedOnBlackHole:
3268 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
3270 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
3272 last = &bq->blocking_queue;
3273 for (t = bq->blocking_queue;
3275 last = &t->link, t = t->link) {
3276 if (t == (StgBlockingQueueElement *)tso) {
3277 *last = (StgBlockingQueueElement *)tso->link;
3281 barf("unblockThread (BLACKHOLE): TSO not found");
3284 case BlockedOnException:
3286 StgTSO *target = tso->block_info.tso;
3288 ASSERT(get_itbl(target)->type == TSO);
3290 if (target->what_next == ThreadRelocated) {
3291 target = target->link;
3292 ASSERT(get_itbl(target)->type == TSO);
3295 ASSERT(target->blocked_exceptions != NULL);
3297 last = (StgBlockingQueueElement **)&target->blocked_exceptions;
3298 for (t = (StgBlockingQueueElement *)target->blocked_exceptions;
3300 last = &t->link, t = t->link) {
3301 ASSERT(get_itbl(t)->type == TSO);
3302 if (t == (StgBlockingQueueElement *)tso) {
3303 *last = (StgBlockingQueueElement *)tso->link;
3307 barf("unblockThread (Exception): TSO not found");
3311 case BlockedOnWrite:
3312 #if defined(mingw32_HOST_OS)
3313 case BlockedOnDoProc:
3316 /* take TSO off blocked_queue */
3317 StgBlockingQueueElement *prev = NULL;
3318 for (t = (StgBlockingQueueElement *)blocked_queue_hd; t != END_BQ_QUEUE;
3319 prev = t, t = t->link) {
3320 if (t == (StgBlockingQueueElement *)tso) {
3322 blocked_queue_hd = (StgTSO *)t->link;
3323 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
3324 blocked_queue_tl = END_TSO_QUEUE;
3327 prev->link = t->link;
3328 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
3329 blocked_queue_tl = (StgTSO *)prev;
3335 barf("unblockThread (I/O): TSO not found");
3338 case BlockedOnDelay:
3340 /* take TSO off sleeping_queue */
3341 StgBlockingQueueElement *prev = NULL;
3342 for (t = (StgBlockingQueueElement *)sleeping_queue; t != END_BQ_QUEUE;
3343 prev = t, t = t->link) {
3344 if (t == (StgBlockingQueueElement *)tso) {
3346 sleeping_queue = (StgTSO *)t->link;
3348 prev->link = t->link;
3353 barf("unblockThread (delay): TSO not found");
3357 barf("unblockThread");
3361 tso->link = END_TSO_QUEUE;
3362 tso->why_blocked = NotBlocked;
3363 tso->block_info.closure = NULL;
3364 PUSH_ON_RUN_QUEUE(tso);
3368 unblockThread(StgTSO *tso)
3372 /* To avoid locking unnecessarily. */
3373 if (tso->why_blocked == NotBlocked) {
3377 switch (tso->why_blocked) {
3380 // Be careful: nothing to do here! We tell the scheduler that the thread
3381 // is runnable and we leave it to the stack-walking code to abort the
3382 // transaction while unwinding the stack. We should perhaps have a debugging
3383 // test to make sure that this really happens and that the 'zombie' transaction
3384 // does not get committed.
3388 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
3390 StgTSO *last_tso = END_TSO_QUEUE;
3391 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
3394 for (t = mvar->head; t != END_TSO_QUEUE;
3395 last = &t->link, last_tso = t, t = t->link) {
3398 if (mvar->tail == tso) {
3399 mvar->tail = last_tso;
3404 barf("unblockThread (MVAR): TSO not found");
3407 case BlockedOnBlackHole:
3409 last = &blackhole_queue;
3410 for (t = blackhole_queue; t != END_TSO_QUEUE;
3411 last = &t->link, t = t->link) {
3417 barf("unblockThread (BLACKHOLE): TSO not found");
3420 case BlockedOnException:
3422 StgTSO *target = tso->block_info.tso;
3424 ASSERT(get_itbl(target)->type == TSO);
3426 while (target->what_next == ThreadRelocated) {
3427 target = target->link;
3428 ASSERT(get_itbl(target)->type == TSO);
3431 ASSERT(target->blocked_exceptions != NULL);
3433 last = &target->blocked_exceptions;
3434 for (t = target->blocked_exceptions; t != END_TSO_QUEUE;
3435 last = &t->link, t = t->link) {
3436 ASSERT(get_itbl(t)->type == TSO);
3442 barf("unblockThread (Exception): TSO not found");
3446 case BlockedOnWrite:
3447 #if defined(mingw32_HOST_OS)
3448 case BlockedOnDoProc:
3451 StgTSO *prev = NULL;
3452 for (t = blocked_queue_hd; t != END_TSO_QUEUE;
3453 prev = t, t = t->link) {
3456 blocked_queue_hd = t->link;
3457 if (blocked_queue_tl == t) {
3458 blocked_queue_tl = END_TSO_QUEUE;
3461 prev->link = t->link;
3462 if (blocked_queue_tl == t) {
3463 blocked_queue_tl = prev;
3469 barf("unblockThread (I/O): TSO not found");
3472 case BlockedOnDelay:
3474 StgTSO *prev = NULL;
3475 for (t = sleeping_queue; t != END_TSO_QUEUE;
3476 prev = t, t = t->link) {
3479 sleeping_queue = t->link;
3481 prev->link = t->link;
3486 barf("unblockThread (delay): TSO not found");
3490 barf("unblockThread");
3494 tso->link = END_TSO_QUEUE;
3495 tso->why_blocked = NotBlocked;
3496 tso->block_info.closure = NULL;
3497 APPEND_TO_RUN_QUEUE(tso);
3501 /* -----------------------------------------------------------------------------
3504 * Check the blackhole_queue for threads that can be woken up. We do
3505 * this periodically: before every GC, and whenever the run queue is
3508 * An elegant solution might be to just wake up all the blocked
3509 * threads with awakenBlockedQueue occasionally: they'll go back to
3510 * sleep again if the object is still a BLACKHOLE. Unfortunately this
3511 * doesn't give us a way to tell whether we've actually managed to
3512 * wake up any threads, so we would be busy-waiting.
3514 * -------------------------------------------------------------------------- */
3517 checkBlackHoles( void )
3520 rtsBool any_woke_up = rtsFalse;
3523 IF_DEBUG(scheduler, sched_belch("checking threads blocked on black holes"));
3525 // ASSUMES: sched_mutex
3526 prev = &blackhole_queue;
3527 t = blackhole_queue;
3528 while (t != END_TSO_QUEUE) {
3529 ASSERT(t->why_blocked == BlockedOnBlackHole);
3530 type = get_itbl(t->block_info.closure)->type;
3531 if (type != BLACKHOLE && type != CAF_BLACKHOLE) {
3532 t = unblockOneLocked(t);
3534 any_woke_up = rtsTrue;
3544 /* -----------------------------------------------------------------------------
3547 * The following function implements the magic for raising an
3548 * asynchronous exception in an existing thread.
3550 * We first remove the thread from any queue on which it might be
3551 * blocked. The possible blockages are MVARs and BLACKHOLE_BQs.
3553 * We strip the stack down to the innermost CATCH_FRAME, building
3554 * thunks in the heap for all the active computations, so they can
3555 * be restarted if necessary. When we reach a CATCH_FRAME, we build
3556 * an application of the handler to the exception, and push it on
3557 * the top of the stack.
3559 * How exactly do we save all the active computations? We create an
3560 * AP_STACK for every UpdateFrame on the stack. Entering one of these
3561 * AP_STACKs pushes everything from the corresponding update frame
3562 * upwards onto the stack. (Actually, it pushes everything up to the
3563 * next update frame plus a pointer to the next AP_STACK object.
3564 * Entering the next AP_STACK object pushes more onto the stack until we
3565 * reach the last AP_STACK object - at which point the stack should look
3566 * exactly as it did when we killed the TSO and we can continue
3567 * execution by entering the closure on top of the stack.
3569 * We can also kill a thread entirely - this happens if either (a) the
3570 * exception passed to raiseAsync is NULL, or (b) there's no
3571 * CATCH_FRAME on the stack. In either case, we strip the entire
3572 * stack and replace the thread with a zombie.
3574 * Locks: sched_mutex held upon entry nor exit.
3576 * -------------------------------------------------------------------------- */
3579 deleteThread(StgTSO *tso)
3581 if (tso->why_blocked != BlockedOnCCall &&
3582 tso->why_blocked != BlockedOnCCall_NoUnblockExc) {
3583 raiseAsync(tso,NULL);
3587 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
3589 deleteThreadImmediately(StgTSO *tso)
3590 { // for forkProcess only:
3591 // delete thread without giving it a chance to catch the KillThread exception
3593 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
3597 if (tso->why_blocked != BlockedOnCCall &&
3598 tso->why_blocked != BlockedOnCCall_NoUnblockExc) {
3602 tso->what_next = ThreadKilled;
3607 raiseAsyncWithLock(StgTSO *tso, StgClosure *exception)
3609 /* When raising async exs from contexts where sched_mutex isn't held;
3610 use raiseAsyncWithLock(). */
3611 ACQUIRE_LOCK(&sched_mutex);
3612 raiseAsync(tso,exception);
3613 RELEASE_LOCK(&sched_mutex);
3617 raiseAsync(StgTSO *tso, StgClosure *exception)
3619 raiseAsync_(tso, exception, rtsFalse);
3623 raiseAsync_(StgTSO *tso, StgClosure *exception, rtsBool stop_at_atomically)
3625 StgRetInfoTable *info;
3628 // Thread already dead?
3629 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
3634 sched_belch("raising exception in thread %ld.", (long)tso->id));
3636 // Remove it from any blocking queues
3641 // The stack freezing code assumes there's a closure pointer on
3642 // the top of the stack, so we have to arrange that this is the case...
3644 if (sp[0] == (W_)&stg_enter_info) {
3648 sp[0] = (W_)&stg_dummy_ret_closure;
3654 // 1. Let the top of the stack be the "current closure"
3656 // 2. Walk up the stack until we find either an UPDATE_FRAME or a
3659 // 3. If it's an UPDATE_FRAME, then make an AP_STACK containing the
3660 // current closure applied to the chunk of stack up to (but not
3661 // including) the update frame. This closure becomes the "current
3662 // closure". Go back to step 2.
3664 // 4. If it's a CATCH_FRAME, then leave the exception handler on
3665 // top of the stack applied to the exception.
3667 // 5. If it's a STOP_FRAME, then kill the thread.
3669 // NB: if we pass an ATOMICALLY_FRAME then abort the associated
3676 info = get_ret_itbl((StgClosure *)frame);
3678 while (info->i.type != UPDATE_FRAME
3679 && (info->i.type != CATCH_FRAME || exception == NULL)
3680 && info->i.type != STOP_FRAME
3681 && (info->i.type != ATOMICALLY_FRAME || stop_at_atomically == rtsFalse))
3683 if (info->i.type == CATCH_RETRY_FRAME || info->i.type == ATOMICALLY_FRAME) {
3684 // IF we find an ATOMICALLY_FRAME then we abort the
3685 // current transaction and propagate the exception. In
3686 // this case (unlike ordinary exceptions) we do not care
3687 // whether the transaction is valid or not because its
3688 // possible validity cannot have caused the exception
3689 // and will not be visible after the abort.
3691 debugBelch("Found atomically block delivering async exception\n"));
3692 stmAbortTransaction(tso -> trec);
3693 tso -> trec = stmGetEnclosingTRec(tso -> trec);
3695 frame += stack_frame_sizeW((StgClosure *)frame);
3696 info = get_ret_itbl((StgClosure *)frame);
3699 switch (info->i.type) {
3701 case ATOMICALLY_FRAME:
3702 ASSERT(stop_at_atomically);
3703 ASSERT(stmGetEnclosingTRec(tso->trec) == NO_TREC);
3704 stmCondemnTransaction(tso -> trec);
3708 // R1 is not a register: the return convention for IO in
3709 // this case puts the return value on the stack, so we
3710 // need to set up the stack to return to the atomically
3711 // frame properly...
3712 tso->sp = frame - 2;
3713 tso->sp[1] = (StgWord) &stg_NO_FINALIZER_closure; // why not?
3714 tso->sp[0] = (StgWord) &stg_ut_1_0_unreg_info;
3716 tso->what_next = ThreadRunGHC;
3720 // If we find a CATCH_FRAME, and we've got an exception to raise,
3721 // then build the THUNK raise(exception), and leave it on
3722 // top of the CATCH_FRAME ready to enter.
3726 StgCatchFrame *cf = (StgCatchFrame *)frame;
3730 // we've got an exception to raise, so let's pass it to the
3731 // handler in this frame.
3733 raise = (StgClosure *)allocate(sizeofW(StgClosure)+1);
3734 TICK_ALLOC_SE_THK(1,0);
3735 SET_HDR(raise,&stg_raise_info,cf->header.prof.ccs);
3736 raise->payload[0] = exception;
3738 // throw away the stack from Sp up to the CATCH_FRAME.
3742 /* Ensure that async excpetions are blocked now, so we don't get
3743 * a surprise exception before we get around to executing the
3746 if (tso->blocked_exceptions == NULL) {
3747 tso->blocked_exceptions = END_TSO_QUEUE;
3750 /* Put the newly-built THUNK on top of the stack, ready to execute
3751 * when the thread restarts.
3754 sp[-1] = (W_)&stg_enter_info;
3756 tso->what_next = ThreadRunGHC;
3757 IF_DEBUG(sanity, checkTSO(tso));
3766 // First build an AP_STACK consisting of the stack chunk above the
3767 // current update frame, with the top word on the stack as the
3770 words = frame - sp - 1;
3771 ap = (StgAP_STACK *)allocate(PAP_sizeW(words));
3774 ap->fun = (StgClosure *)sp[0];
3776 for(i=0; i < (nat)words; ++i) {
3777 ap->payload[i] = (StgClosure *)*sp++;
3780 SET_HDR(ap,&stg_AP_STACK_info,
3781 ((StgClosure *)frame)->header.prof.ccs /* ToDo */);
3782 TICK_ALLOC_UP_THK(words+1,0);
3785 debugBelch("sched: Updating ");
3786 printPtr((P_)((StgUpdateFrame *)frame)->updatee);
3787 debugBelch(" with ");
3788 printObj((StgClosure *)ap);
3791 // Replace the updatee with an indirection - happily
3792 // this will also wake up any threads currently
3793 // waiting on the result.
3795 // Warning: if we're in a loop, more than one update frame on
3796 // the stack may point to the same object. Be careful not to
3797 // overwrite an IND_OLDGEN in this case, because we'll screw
3798 // up the mutable lists. To be on the safe side, don't
3799 // overwrite any kind of indirection at all. See also
3800 // threadSqueezeStack in GC.c, where we have to make a similar
3803 if (!closure_IND(((StgUpdateFrame *)frame)->updatee)) {
3804 // revert the black hole
3805 UPD_IND_NOLOCK(((StgUpdateFrame *)frame)->updatee,
3808 sp += sizeofW(StgUpdateFrame) - 1;
3809 sp[0] = (W_)ap; // push onto stack
3814 // We've stripped the entire stack, the thread is now dead.
3815 sp += sizeofW(StgStopFrame);
3816 tso->what_next = ThreadKilled;
3827 /* -----------------------------------------------------------------------------
3828 raiseExceptionHelper
3830 This function is called by the raise# primitve, just so that we can
3831 move some of the tricky bits of raising an exception from C-- into
3832 C. Who knows, it might be a useful re-useable thing here too.
3833 -------------------------------------------------------------------------- */
3836 raiseExceptionHelper (StgTSO *tso, StgClosure *exception)
3838 StgClosure *raise_closure = NULL;
3840 StgRetInfoTable *info;
3842 // This closure represents the expression 'raise# E' where E
3843 // is the exception raise. It is used to overwrite all the
3844 // thunks which are currently under evaluataion.
3848 // LDV profiling: stg_raise_info has THUNK as its closure
3849 // type. Since a THUNK takes at least MIN_UPD_SIZE words in its
3850 // payload, MIN_UPD_SIZE is more approprate than 1. It seems that
3851 // 1 does not cause any problem unless profiling is performed.
3852 // However, when LDV profiling goes on, we need to linearly scan
3853 // small object pool, where raise_closure is stored, so we should
3854 // use MIN_UPD_SIZE.
3856 // raise_closure = (StgClosure *)RET_STGCALL1(P_,allocate,
3857 // sizeofW(StgClosure)+1);
3861 // Walk up the stack, looking for the catch frame. On the way,
3862 // we update any closures pointed to from update frames with the
3863 // raise closure that we just built.
3867 info = get_ret_itbl((StgClosure *)p);
3868 next = p + stack_frame_sizeW((StgClosure *)p);
3869 switch (info->i.type) {
3872 // Only create raise_closure if we need to.
3873 if (raise_closure == NULL) {
3875 (StgClosure *)allocate(sizeofW(StgClosure)+MIN_UPD_SIZE);
3876 SET_HDR(raise_closure, &stg_raise_info, CCCS);
3877 raise_closure->payload[0] = exception;
3879 UPD_IND(((StgUpdateFrame *)p)->updatee,raise_closure);
3883 case ATOMICALLY_FRAME:
3884 IF_DEBUG(stm, debugBelch("Found ATOMICALLY_FRAME at %p\n", p));
3886 return ATOMICALLY_FRAME;
3892 case CATCH_STM_FRAME:
3893 IF_DEBUG(stm, debugBelch("Found CATCH_STM_FRAME at %p\n", p));
3895 return CATCH_STM_FRAME;
3901 case CATCH_RETRY_FRAME:
3910 /* -----------------------------------------------------------------------------
3911 findRetryFrameHelper
3913 This function is called by the retry# primitive. It traverses the stack
3914 leaving tso->sp referring to the frame which should handle the retry.
3916 This should either be a CATCH_RETRY_FRAME (if the retry# is within an orElse#)
3917 or should be a ATOMICALLY_FRAME (if the retry# reaches the top level).
3919 We skip CATCH_STM_FRAMEs because retries are not considered to be exceptions,
3920 despite the similar implementation.
3922 We should not expect to see CATCH_FRAME or STOP_FRAME because those should
3923 not be created within memory transactions.
3924 -------------------------------------------------------------------------- */
3927 findRetryFrameHelper (StgTSO *tso)
3930 StgRetInfoTable *info;
3934 info = get_ret_itbl((StgClosure *)p);
3935 next = p + stack_frame_sizeW((StgClosure *)p);
3936 switch (info->i.type) {
3938 case ATOMICALLY_FRAME:
3939 IF_DEBUG(stm, debugBelch("Found ATOMICALLY_FRAME at %p during retrry\n", p));
3941 return ATOMICALLY_FRAME;
3943 case CATCH_RETRY_FRAME:
3944 IF_DEBUG(stm, debugBelch("Found CATCH_RETRY_FRAME at %p during retrry\n", p));
3946 return CATCH_RETRY_FRAME;
3948 case CATCH_STM_FRAME:
3950 ASSERT(info->i.type != CATCH_FRAME);
3951 ASSERT(info->i.type != STOP_FRAME);
3958 /* -----------------------------------------------------------------------------
3959 resurrectThreads is called after garbage collection on the list of
3960 threads found to be garbage. Each of these threads will be woken
3961 up and sent a signal: BlockedOnDeadMVar if the thread was blocked
3962 on an MVar, or NonTermination if the thread was blocked on a Black
3965 Locks: sched_mutex isn't held upon entry nor exit.
3966 -------------------------------------------------------------------------- */
3969 resurrectThreads( StgTSO *threads )
3973 for (tso = threads; tso != END_TSO_QUEUE; tso = next) {
3974 next = tso->global_link;
3975 tso->global_link = all_threads;
3977 IF_DEBUG(scheduler, sched_belch("resurrecting thread %d", tso->id));
3979 switch (tso->why_blocked) {
3981 case BlockedOnException:
3982 /* Called by GC - sched_mutex lock is currently held. */
3983 raiseAsync(tso,(StgClosure *)BlockedOnDeadMVar_closure);
3985 case BlockedOnBlackHole:
3986 raiseAsync(tso,(StgClosure *)NonTermination_closure);
3989 raiseAsync(tso,(StgClosure *)BlockedIndefinitely_closure);
3992 /* This might happen if the thread was blocked on a black hole
3993 * belonging to a thread that we've just woken up (raiseAsync
3994 * can wake up threads, remember...).
3998 barf("resurrectThreads: thread blocked in a strange way");
4003 /* ----------------------------------------------------------------------------
4004 * Debugging: why is a thread blocked
4005 * [Also provides useful information when debugging threaded programs
4006 * at the Haskell source code level, so enable outside of DEBUG. --sof 7/02]
4007 ------------------------------------------------------------------------- */
4010 printThreadBlockage(StgTSO *tso)
4012 switch (tso->why_blocked) {
4014 debugBelch("is blocked on read from fd %ld", tso->block_info.fd);
4016 case BlockedOnWrite:
4017 debugBelch("is blocked on write to fd %ld", tso->block_info.fd);
4019 #if defined(mingw32_HOST_OS)
4020 case BlockedOnDoProc:
4021 debugBelch("is blocked on proc (request: %ld)", tso->block_info.async_result->reqID);
4024 case BlockedOnDelay:
4025 debugBelch("is blocked until %ld", tso->block_info.target);
4028 debugBelch("is blocked on an MVar");
4030 case BlockedOnException:
4031 debugBelch("is blocked on delivering an exception to thread %d",
4032 tso->block_info.tso->id);
4034 case BlockedOnBlackHole:
4035 debugBelch("is blocked on a black hole");
4038 debugBelch("is not blocked");
4040 #if defined(PARALLEL_HASKELL)
4042 debugBelch("is blocked on global address; local FM_BQ is %p (%s)",
4043 tso->block_info.closure, info_type(tso->block_info.closure));
4045 case BlockedOnGA_NoSend:
4046 debugBelch("is blocked on global address (no send); local FM_BQ is %p (%s)",
4047 tso->block_info.closure, info_type(tso->block_info.closure));
4050 case BlockedOnCCall:
4051 debugBelch("is blocked on an external call");
4053 case BlockedOnCCall_NoUnblockExc:
4054 debugBelch("is blocked on an external call (exceptions were already blocked)");
4057 debugBelch("is blocked on an STM operation");
4060 barf("printThreadBlockage: strange tso->why_blocked: %d for TSO %d (%d)",
4061 tso->why_blocked, tso->id, tso);
4066 printThreadStatus(StgTSO *tso)
4068 switch (tso->what_next) {
4070 debugBelch("has been killed");
4072 case ThreadComplete:
4073 debugBelch("has completed");
4076 printThreadBlockage(tso);
4081 printAllThreads(void)
4086 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
4087 ullong_format_string(TIME_ON_PROC(CurrentProc),
4088 time_string, rtsFalse/*no commas!*/);
4090 debugBelch("all threads at [%s]:\n", time_string);
4091 # elif defined(PARALLEL_HASKELL)
4092 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
4093 ullong_format_string(CURRENT_TIME,
4094 time_string, rtsFalse/*no commas!*/);
4096 debugBelch("all threads at [%s]:\n", time_string);
4098 debugBelch("all threads:\n");
4101 for (t = all_threads; t != END_TSO_QUEUE; t = t->global_link) {
4102 debugBelch("\tthread %d @ %p ", t->id, (void *)t);
4105 void *label = lookupThreadLabel(t->id);
4106 if (label) debugBelch("[\"%s\"] ",(char *)label);
4109 printThreadStatus(t);
4117 Print a whole blocking queue attached to node (debugging only).
4119 # if defined(PARALLEL_HASKELL)
4121 print_bq (StgClosure *node)
4123 StgBlockingQueueElement *bqe;
4127 debugBelch("## BQ of closure %p (%s): ",
4128 node, info_type(node));
4130 /* should cover all closures that may have a blocking queue */
4131 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
4132 get_itbl(node)->type == FETCH_ME_BQ ||
4133 get_itbl(node)->type == RBH ||
4134 get_itbl(node)->type == MVAR);
4136 ASSERT(node!=(StgClosure*)NULL); // sanity check
4138 print_bqe(((StgBlockingQueue*)node)->blocking_queue);
4142 Print a whole blocking queue starting with the element bqe.
4145 print_bqe (StgBlockingQueueElement *bqe)
4150 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
4152 for (end = (bqe==END_BQ_QUEUE);
4153 !end; // iterate until bqe points to a CONSTR
4154 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE),
4155 bqe = end ? END_BQ_QUEUE : bqe->link) {
4156 ASSERT(bqe != END_BQ_QUEUE); // sanity check
4157 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
4158 /* types of closures that may appear in a blocking queue */
4159 ASSERT(get_itbl(bqe)->type == TSO ||
4160 get_itbl(bqe)->type == BLOCKED_FETCH ||
4161 get_itbl(bqe)->type == CONSTR);
4162 /* only BQs of an RBH end with an RBH_Save closure */
4163 //ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
4165 switch (get_itbl(bqe)->type) {
4167 debugBelch(" TSO %u (%x),",
4168 ((StgTSO *)bqe)->id, ((StgTSO *)bqe));
4171 debugBelch(" BF (node=%p, ga=((%x, %d, %x)),",
4172 ((StgBlockedFetch *)bqe)->node,
4173 ((StgBlockedFetch *)bqe)->ga.payload.gc.gtid,
4174 ((StgBlockedFetch *)bqe)->ga.payload.gc.slot,
4175 ((StgBlockedFetch *)bqe)->ga.weight);
4178 debugBelch(" %s (IP %p),",
4179 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
4180 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
4181 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
4182 "RBH_Save_?"), get_itbl(bqe));
4185 barf("Unexpected closure type %s in blocking queue", // of %p (%s)",
4186 info_type((StgClosure *)bqe)); // , node, info_type(node));
4192 # elif defined(GRAN)
4194 print_bq (StgClosure *node)
4196 StgBlockingQueueElement *bqe;
4197 PEs node_loc, tso_loc;
4200 /* should cover all closures that may have a blocking queue */
4201 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
4202 get_itbl(node)->type == FETCH_ME_BQ ||
4203 get_itbl(node)->type == RBH);
4205 ASSERT(node!=(StgClosure*)NULL); // sanity check
4206 node_loc = where_is(node);
4208 debugBelch("## BQ of closure %p (%s) on [PE %d]: ",
4209 node, info_type(node), node_loc);
4212 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
4214 for (bqe = ((StgBlockingQueue*)node)->blocking_queue, end = (bqe==END_BQ_QUEUE);
4215 !end; // iterate until bqe points to a CONSTR
4216 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE), bqe = end ? END_BQ_QUEUE : bqe->link) {
4217 ASSERT(bqe != END_BQ_QUEUE); // sanity check
4218 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
4219 /* types of closures that may appear in a blocking queue */
4220 ASSERT(get_itbl(bqe)->type == TSO ||
4221 get_itbl(bqe)->type == CONSTR);
4222 /* only BQs of an RBH end with an RBH_Save closure */
4223 ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
4225 tso_loc = where_is((StgClosure *)bqe);
4226 switch (get_itbl(bqe)->type) {
4228 debugBelch(" TSO %d (%p) on [PE %d],",
4229 ((StgTSO *)bqe)->id, (StgTSO *)bqe, tso_loc);
4232 debugBelch(" %s (IP %p),",
4233 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
4234 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
4235 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
4236 "RBH_Save_?"), get_itbl(bqe));
4239 barf("Unexpected closure type %s in blocking queue of %p (%s)",
4240 info_type((StgClosure *)bqe), node, info_type(node));
4248 #if defined(PARALLEL_HASKELL)
4255 for (i=0, tso=run_queue_hd;
4256 tso != END_TSO_QUEUE;
4265 sched_belch(char *s, ...)
4269 #ifdef RTS_SUPPORTS_THREADS
4270 debugBelch("sched (task %p): ", osThreadId());
4271 #elif defined(PARALLEL_HASKELL)
4274 debugBelch("sched: ");