1 /* ---------------------------------------------------------------------------
2 * $Id: Schedule.c,v 1.196 2004/05/06 12:20:04 wolfgang Exp $
4 * (c) The GHC Team, 1998-2003
8 * Different GHC ways use this scheduler quite differently (see comments below)
9 * Here is the global picture:
11 * WAY Name CPP flag What's it for
12 * --------------------------------------
13 * mp GUM PAR Parallel execution on a distrib. memory machine
14 * s SMP SMP Parallel execution on a shared memory machine
15 * mg GranSim GRAN Simulation of parallel execution
16 * md GUM/GdH DIST Distributed execution (based on GUM)
18 * --------------------------------------------------------------------------*/
21 * Version with support for distributed memory parallelism aka GUM (WAY=mp):
23 The main scheduling loop in GUM iterates until a finish message is received.
24 In that case a global flag @receivedFinish@ is set and this instance of
25 the RTS shuts down. See ghc/rts/parallel/HLComms.c:processMessages()
26 for the handling of incoming messages, such as PP_FINISH.
27 Note that in the parallel case we have a system manager that coordinates
28 different PEs, each of which are running one instance of the RTS.
29 See ghc/rts/parallel/SysMan.c for the main routine of the parallel program.
30 From this routine processes executing ghc/rts/Main.c are spawned. -- HWL
32 * Version with support for simulating parallel execution aka GranSim (WAY=mg):
34 The main scheduling code in GranSim is quite different from that in std
35 (concurrent) Haskell: while concurrent Haskell just iterates over the
36 threads in the runnable queue, GranSim is event driven, i.e. it iterates
37 over the events in the global event queue. -- HWL
40 #include "PosixSource.h"
47 #include "StgStartup.h"
49 #define COMPILING_SCHEDULER
51 #include "StgMiscClosures.h"
53 #include "Interpreter.h"
54 #include "Exception.h"
61 #include "ThreadLabels.h"
63 #include "Proftimer.h"
66 #if defined(GRAN) || defined(PAR)
67 # include "GranSimRts.h"
69 # include "ParallelRts.h"
70 # include "Parallel.h"
71 # include "ParallelDebug.h"
76 #include "Capability.h"
77 #include "OSThreads.h"
80 #ifdef HAVE_SYS_TYPES_H
81 #include <sys/types.h>
96 #define USED_IN_THREADED_RTS
98 #define USED_IN_THREADED_RTS STG_UNUSED
101 #ifdef RTS_SUPPORTS_THREADS
102 #define USED_WHEN_RTS_SUPPORTS_THREADS
104 #define USED_WHEN_RTS_SUPPORTS_THREADS STG_UNUSED
107 /* Main thread queue.
108 * Locks required: sched_mutex.
110 StgMainThread *main_threads = NULL;
113 * Locks required: sched_mutex.
117 StgTSO* ActiveTSO = NULL; /* for assigning system costs; GranSim-Light only */
118 /* rtsTime TimeOfNextEvent, EndOfTimeSlice; now in GranSim.c */
121 In GranSim we have a runnable and a blocked queue for each processor.
122 In order to minimise code changes new arrays run_queue_hds/tls
123 are created. run_queue_hd is then a short cut (macro) for
124 run_queue_hds[CurrentProc] (see GranSim.h).
127 StgTSO *run_queue_hds[MAX_PROC], *run_queue_tls[MAX_PROC];
128 StgTSO *blocked_queue_hds[MAX_PROC], *blocked_queue_tls[MAX_PROC];
129 StgTSO *ccalling_threadss[MAX_PROC];
130 /* We use the same global list of threads (all_threads) in GranSim as in
131 the std RTS (i.e. we are cheating). However, we don't use this list in
132 the GranSim specific code at the moment (so we are only potentially
137 StgTSO *run_queue_hd = NULL;
138 StgTSO *run_queue_tl = NULL;
139 StgTSO *blocked_queue_hd = NULL;
140 StgTSO *blocked_queue_tl = NULL;
141 StgTSO *sleeping_queue = NULL; /* perhaps replace with a hash table? */
145 /* Linked list of all threads.
146 * Used for detecting garbage collected threads.
148 StgTSO *all_threads = NULL;
150 /* When a thread performs a safe C call (_ccall_GC, using old
151 * terminology), it gets put on the suspended_ccalling_threads
152 * list. Used by the garbage collector.
154 static StgTSO *suspended_ccalling_threads;
156 static StgTSO *threadStackOverflow(StgTSO *tso);
158 /* KH: The following two flags are shared memory locations. There is no need
159 to lock them, since they are only unset at the end of a scheduler
163 /* flag set by signal handler to precipitate a context switch */
164 nat context_switch = 0;
166 /* if this flag is set as well, give up execution */
167 rtsBool interrupted = rtsFalse;
169 /* Next thread ID to allocate.
170 * Locks required: thread_id_mutex
172 static StgThreadID next_thread_id = 1;
175 * Pointers to the state of the current thread.
176 * Rule of thumb: if CurrentTSO != NULL, then we're running a Haskell
177 * thread. If CurrentTSO == NULL, then we're at the scheduler level.
180 /* The smallest stack size that makes any sense is:
181 * RESERVED_STACK_WORDS (so we can get back from the stack overflow)
182 * + sizeofW(StgStopFrame) (the stg_stop_thread_info frame)
183 * + 1 (the closure to enter)
185 * + 1 (spare slot req'd by stg_ap_v_ret)
187 * A thread with this stack will bomb immediately with a stack
188 * overflow, which will increase its stack size.
191 #define MIN_STACK_WORDS (RESERVED_STACK_WORDS + sizeofW(StgStopFrame) + 3)
198 /* This is used in `TSO.h' and gcc 2.96 insists that this variable actually
199 * exists - earlier gccs apparently didn't.
204 static rtsBool ready_to_gc;
207 * Set to TRUE when entering a shutdown state (via shutdownHaskellAndExit()) --
208 * in an MT setting, needed to signal that a worker thread shouldn't hang around
209 * in the scheduler when it is out of work.
211 static rtsBool shutting_down_scheduler = rtsFalse;
213 void addToBlockedQueue ( StgTSO *tso );
215 static void schedule ( StgMainThread *mainThread, Capability *initialCapability );
216 void interruptStgRts ( void );
218 static void detectBlackHoles ( void );
220 #if defined(RTS_SUPPORTS_THREADS)
221 /* ToDo: carefully document the invariants that go together
222 * with these synchronisation objects.
224 Mutex sched_mutex = INIT_MUTEX_VAR;
225 Mutex term_mutex = INIT_MUTEX_VAR;
227 #endif /* RTS_SUPPORTS_THREADS */
231 rtsTime TimeOfLastYield;
232 rtsBool emitSchedule = rtsTrue;
236 static char *whatNext_strs[] = {
246 StgTSO * createSparkThread(rtsSpark spark);
247 StgTSO * activateSpark (rtsSpark spark);
250 /* ----------------------------------------------------------------------------
252 * ------------------------------------------------------------------------- */
254 #if defined(RTS_SUPPORTS_THREADS)
255 static rtsBool startingWorkerThread = rtsFalse;
257 static void taskStart(void);
261 ACQUIRE_LOCK(&sched_mutex);
262 startingWorkerThread = rtsFalse;
264 RELEASE_LOCK(&sched_mutex);
268 startSchedulerTaskIfNecessary(void)
270 if(run_queue_hd != END_TSO_QUEUE
271 || blocked_queue_hd != END_TSO_QUEUE
272 || sleeping_queue != END_TSO_QUEUE)
274 if(!startingWorkerThread)
275 { // we don't want to start another worker thread
276 // just because the last one hasn't yet reached the
277 // "waiting for capability" state
278 startingWorkerThread = rtsTrue;
279 if(!startTask(taskStart))
281 startingWorkerThread = rtsFalse;
288 /* ---------------------------------------------------------------------------
289 Main scheduling loop.
291 We use round-robin scheduling, each thread returning to the
292 scheduler loop when one of these conditions is detected:
295 * timer expires (thread yields)
300 Locking notes: we acquire the scheduler lock once at the beginning
301 of the scheduler loop, and release it when
303 * running a thread, or
304 * waiting for work, or
305 * waiting for a GC to complete.
308 In a GranSim setup this loop iterates over the global event queue.
309 This revolves around the global event queue, which determines what
310 to do next. Therefore, it's more complicated than either the
311 concurrent or the parallel (GUM) setup.
314 GUM iterates over incoming messages.
315 It starts with nothing to do (thus CurrentTSO == END_TSO_QUEUE),
316 and sends out a fish whenever it has nothing to do; in-between
317 doing the actual reductions (shared code below) it processes the
318 incoming messages and deals with delayed operations
319 (see PendingFetches).
320 This is not the ugliest code you could imagine, but it's bloody close.
322 ------------------------------------------------------------------------ */
324 schedule( StgMainThread *mainThread USED_WHEN_RTS_SUPPORTS_THREADS,
325 Capability *initialCapability )
329 StgThreadReturnCode ret;
337 rtsBool receivedFinish = rtsFalse;
339 nat tp_size, sp_size; // stats only
342 rtsBool was_interrupted = rtsFalse;
343 StgTSOWhatNext prev_what_next;
345 // Pre-condition: sched_mutex is held.
346 // We might have a capability, passed in as initialCapability.
347 cap = initialCapability;
349 #if defined(RTS_SUPPORTS_THREADS)
351 // in the threaded case, the capability is either passed in via the
352 // initialCapability parameter, or initialized inside the scheduler
356 sched_belch("### NEW SCHEDULER LOOP (main thr: %p, cap: %p)",
357 mainThread, initialCapability);
360 // simply initialise it in the non-threaded case
361 grabCapability(&cap);
365 /* set up first event to get things going */
366 /* ToDo: assign costs for system setup and init MainTSO ! */
367 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
369 CurrentTSO, (StgClosure*)NULL, (rtsSpark*)NULL);
372 fprintf(stderr, "GRAN: Init CurrentTSO (in schedule) = %p\n", CurrentTSO);
373 G_TSO(CurrentTSO, 5));
375 if (RtsFlags.GranFlags.Light) {
376 /* Save current time; GranSim Light only */
377 CurrentTSO->gran.clock = CurrentTime[CurrentProc];
380 event = get_next_event();
382 while (event!=(rtsEvent*)NULL) {
383 /* Choose the processor with the next event */
384 CurrentProc = event->proc;
385 CurrentTSO = event->tso;
389 while (!receivedFinish) { /* set by processMessages */
390 /* when receiving PP_FINISH message */
392 #else // everything except GRAN and PAR
398 IF_DEBUG(scheduler, printAllThreads());
400 #if defined(RTS_SUPPORTS_THREADS)
401 // Yield the capability to higher-priority tasks if necessary.
404 yieldCapability(&cap);
407 // If we do not currently hold a capability, we wait for one
410 waitForCapability(&sched_mutex, &cap,
411 mainThread ? &mainThread->bound_thread_cond : NULL);
414 // We now have a capability...
418 // If we're interrupted (the user pressed ^C, or some other
419 // termination condition occurred), kill all the currently running
423 IF_DEBUG(scheduler, sched_belch("interrupted"));
424 interrupted = rtsFalse;
425 was_interrupted = rtsTrue;
426 #if defined(RTS_SUPPORTS_THREADS)
427 // In the threaded RTS, deadlock detection doesn't work,
428 // so just exit right away.
429 prog_belch("interrupted");
430 releaseCapability(cap);
431 RELEASE_LOCK(&sched_mutex);
432 shutdownHaskellAndExit(EXIT_SUCCESS);
438 #if defined(RTS_USER_SIGNALS)
439 // check for signals each time around the scheduler
440 if (signals_pending()) {
441 RELEASE_LOCK(&sched_mutex); /* ToDo: kill */
442 startSignalHandlers();
443 ACQUIRE_LOCK(&sched_mutex);
448 // Check whether any waiting threads need to be woken up. If the
449 // run queue is empty, and there are no other tasks running, we
450 // can wait indefinitely for something to happen.
452 if ( !EMPTY_QUEUE(blocked_queue_hd) || !EMPTY_QUEUE(sleeping_queue)
453 #if defined(RTS_SUPPORTS_THREADS)
458 awaitEvent( EMPTY_RUN_QUEUE() );
460 // we can be interrupted while waiting for I/O...
461 if (interrupted) continue;
464 * Detect deadlock: when we have no threads to run, there are no
465 * threads waiting on I/O or sleeping, and all the other tasks are
466 * waiting for work, we must have a deadlock of some description.
468 * We first try to find threads blocked on themselves (ie. black
469 * holes), and generate NonTermination exceptions where necessary.
471 * If no threads are black holed, we have a deadlock situation, so
472 * inform all the main threads.
474 #if !defined(PAR) && !defined(RTS_SUPPORTS_THREADS)
475 if ( EMPTY_THREAD_QUEUES() )
477 IF_DEBUG(scheduler, sched_belch("deadlocked, forcing major GC..."));
478 // Garbage collection can release some new threads due to
479 // either (a) finalizers or (b) threads resurrected because
480 // they are about to be send BlockedOnDeadMVar. Any threads
481 // thus released will be immediately runnable.
482 GarbageCollect(GetRoots,rtsTrue);
484 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
487 sched_belch("still deadlocked, checking for black holes..."));
490 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
492 #if defined(RTS_USER_SIGNALS)
493 /* If we have user-installed signal handlers, then wait
494 * for signals to arrive rather then bombing out with a
497 if ( anyUserHandlers() ) {
499 sched_belch("still deadlocked, waiting for signals..."));
503 // we might be interrupted...
504 if (interrupted) { continue; }
506 if (signals_pending()) {
507 RELEASE_LOCK(&sched_mutex);
508 startSignalHandlers();
509 ACQUIRE_LOCK(&sched_mutex);
511 ASSERT(!EMPTY_RUN_QUEUE());
516 /* Probably a real deadlock. Send the current main thread the
517 * Deadlock exception (or in the SMP build, send *all* main
518 * threads the deadlock exception, since none of them can make
524 switch (m->tso->why_blocked) {
525 case BlockedOnBlackHole:
526 case BlockedOnException:
528 raiseAsync(m->tso, (StgClosure *)NonTermination_closure);
531 barf("deadlock: main thread blocked in a strange way");
537 #elif defined(RTS_SUPPORTS_THREADS)
538 // ToDo: add deadlock detection in threaded RTS
540 // ToDo: add deadlock detection in GUM (similar to SMP) -- HWL
543 #if defined(RTS_SUPPORTS_THREADS)
544 if ( EMPTY_RUN_QUEUE() ) {
545 continue; // nothing to do
550 if (RtsFlags.GranFlags.Light)
551 GranSimLight_enter_system(event, &ActiveTSO); // adjust ActiveTSO etc
553 /* adjust time based on time-stamp */
554 if (event->time > CurrentTime[CurrentProc] &&
555 event->evttype != ContinueThread)
556 CurrentTime[CurrentProc] = event->time;
558 /* Deal with the idle PEs (may issue FindWork or MoveSpark events) */
559 if (!RtsFlags.GranFlags.Light)
562 IF_DEBUG(gran, fprintf(stderr, "GRAN: switch by event-type\n"));
564 /* main event dispatcher in GranSim */
565 switch (event->evttype) {
566 /* Should just be continuing execution */
568 IF_DEBUG(gran, fprintf(stderr, "GRAN: doing ContinueThread\n"));
569 /* ToDo: check assertion
570 ASSERT(run_queue_hd != (StgTSO*)NULL &&
571 run_queue_hd != END_TSO_QUEUE);
573 /* Ignore ContinueThreads for fetching threads (if synchr comm) */
574 if (!RtsFlags.GranFlags.DoAsyncFetch &&
575 procStatus[CurrentProc]==Fetching) {
576 belch("ghuH: Spurious ContinueThread while Fetching ignored; TSO %d (%p) [PE %d]",
577 CurrentTSO->id, CurrentTSO, CurrentProc);
580 /* Ignore ContinueThreads for completed threads */
581 if (CurrentTSO->what_next == ThreadComplete) {
582 belch("ghuH: found a ContinueThread event for completed thread %d (%p) [PE %d] (ignoring ContinueThread)",
583 CurrentTSO->id, CurrentTSO, CurrentProc);
586 /* Ignore ContinueThreads for threads that are being migrated */
587 if (PROCS(CurrentTSO)==Nowhere) {
588 belch("ghuH: trying to run the migrating TSO %d (%p) [PE %d] (ignoring ContinueThread)",
589 CurrentTSO->id, CurrentTSO, CurrentProc);
592 /* The thread should be at the beginning of the run queue */
593 if (CurrentTSO!=run_queue_hds[CurrentProc]) {
594 belch("ghuH: TSO %d (%p) [PE %d] is not at the start of the run_queue when doing a ContinueThread",
595 CurrentTSO->id, CurrentTSO, CurrentProc);
596 break; // run the thread anyway
599 new_event(proc, proc, CurrentTime[proc],
601 (StgTSO*)NULL, (StgClosure*)NULL, (rtsSpark*)NULL);
603 */ /* Catches superfluous CONTINUEs -- should be unnecessary */
604 break; // now actually run the thread; DaH Qu'vam yImuHbej
607 do_the_fetchnode(event);
608 goto next_thread; /* handle next event in event queue */
611 do_the_globalblock(event);
612 goto next_thread; /* handle next event in event queue */
615 do_the_fetchreply(event);
616 goto next_thread; /* handle next event in event queue */
618 case UnblockThread: /* Move from the blocked queue to the tail of */
619 do_the_unblock(event);
620 goto next_thread; /* handle next event in event queue */
622 case ResumeThread: /* Move from the blocked queue to the tail of */
623 /* the runnable queue ( i.e. Qu' SImqa'lu') */
624 event->tso->gran.blocktime +=
625 CurrentTime[CurrentProc] - event->tso->gran.blockedat;
626 do_the_startthread(event);
627 goto next_thread; /* handle next event in event queue */
630 do_the_startthread(event);
631 goto next_thread; /* handle next event in event queue */
634 do_the_movethread(event);
635 goto next_thread; /* handle next event in event queue */
638 do_the_movespark(event);
639 goto next_thread; /* handle next event in event queue */
642 do_the_findwork(event);
643 goto next_thread; /* handle next event in event queue */
646 barf("Illegal event type %u\n", event->evttype);
649 /* This point was scheduler_loop in the old RTS */
651 IF_DEBUG(gran, belch("GRAN: after main switch"));
653 TimeOfLastEvent = CurrentTime[CurrentProc];
654 TimeOfNextEvent = get_time_of_next_event();
655 IgnoreEvents=(TimeOfNextEvent==0); // HWL HACK
656 // CurrentTSO = ThreadQueueHd;
658 IF_DEBUG(gran, belch("GRAN: time of next event is: %ld",
661 if (RtsFlags.GranFlags.Light)
662 GranSimLight_leave_system(event, &ActiveTSO);
664 EndOfTimeSlice = CurrentTime[CurrentProc]+RtsFlags.GranFlags.time_slice;
667 belch("GRAN: end of time-slice is %#lx", EndOfTimeSlice));
669 /* in a GranSim setup the TSO stays on the run queue */
671 /* Take a thread from the run queue. */
672 POP_RUN_QUEUE(t); // take_off_run_queue(t);
675 fprintf(stderr, "GRAN: About to run current thread, which is\n");
678 context_switch = 0; // turned on via GranYield, checking events and time slice
681 DumpGranEvent(GR_SCHEDULE, t));
683 procStatus[CurrentProc] = Busy;
686 if (PendingFetches != END_BF_QUEUE) {
690 /* ToDo: phps merge with spark activation above */
691 /* check whether we have local work and send requests if we have none */
692 if (EMPTY_RUN_QUEUE()) { /* no runnable threads */
693 /* :-[ no local threads => look out for local sparks */
694 /* the spark pool for the current PE */
695 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
696 if (advisory_thread_count < RtsFlags.ParFlags.maxThreads &&
697 pool->hd < pool->tl) {
699 * ToDo: add GC code check that we really have enough heap afterwards!!
701 * If we're here (no runnable threads) and we have pending
702 * sparks, we must have a space problem. Get enough space
703 * to turn one of those pending sparks into a
707 spark = findSpark(rtsFalse); /* get a spark */
708 if (spark != (rtsSpark) NULL) {
709 tso = activateSpark(spark); /* turn the spark into a thread */
710 IF_PAR_DEBUG(schedule,
711 belch("==== schedule: Created TSO %d (%p); %d threads active",
712 tso->id, tso, advisory_thread_count));
714 if (tso==END_TSO_QUEUE) { /* failed to activate spark->back to loop */
715 belch("==^^ failed to activate spark");
717 } /* otherwise fall through & pick-up new tso */
719 IF_PAR_DEBUG(verbose,
720 belch("==^^ no local sparks (spark pool contains only NFs: %d)",
721 spark_queue_len(pool)));
726 /* If we still have no work we need to send a FISH to get a spark
729 if (EMPTY_RUN_QUEUE()) {
730 /* =8-[ no local sparks => look for work on other PEs */
732 * We really have absolutely no work. Send out a fish
733 * (there may be some out there already), and wait for
734 * something to arrive. We clearly can't run any threads
735 * until a SCHEDULE or RESUME arrives, and so that's what
736 * we're hoping to see. (Of course, we still have to
737 * respond to other types of messages.)
739 TIME now = msTime() /*CURRENT_TIME*/;
740 IF_PAR_DEBUG(verbose,
741 belch("-- now=%ld", now));
742 IF_PAR_DEBUG(verbose,
743 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
744 (last_fish_arrived_at!=0 &&
745 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay > now)) {
746 belch("--$$ delaying FISH until %ld (last fish %ld, delay %ld, now %ld)",
747 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay,
748 last_fish_arrived_at,
749 RtsFlags.ParFlags.fishDelay, now);
752 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
753 (last_fish_arrived_at==0 ||
754 (last_fish_arrived_at+RtsFlags.ParFlags.fishDelay <= now))) {
755 /* outstandingFishes is set in sendFish, processFish;
756 avoid flooding system with fishes via delay */
758 sendFish(pe, mytid, NEW_FISH_AGE, NEW_FISH_HISTORY,
761 // Global statistics: count no. of fishes
762 if (RtsFlags.ParFlags.ParStats.Global &&
763 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
764 globalParStats.tot_fish_mess++;
768 receivedFinish = processMessages();
771 } else if (PacketsWaiting()) { /* Look for incoming messages */
772 receivedFinish = processMessages();
775 /* Now we are sure that we have some work available */
776 ASSERT(run_queue_hd != END_TSO_QUEUE);
778 /* Take a thread from the run queue, if we have work */
779 POP_RUN_QUEUE(t); // take_off_run_queue(END_TSO_QUEUE);
780 IF_DEBUG(sanity,checkTSO(t));
782 /* ToDo: write something to the log-file
783 if (RTSflags.ParFlags.granSimStats && !sameThread)
784 DumpGranEvent(GR_SCHEDULE, RunnableThreadsHd);
788 /* the spark pool for the current PE */
789 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
792 belch("--=^ %d threads, %d sparks on [%#x]",
793 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
796 if (0 && RtsFlags.ParFlags.ParStats.Full &&
797 t && LastTSO && t->id != LastTSO->id &&
798 LastTSO->why_blocked == NotBlocked &&
799 LastTSO->what_next != ThreadComplete) {
800 // if previously scheduled TSO not blocked we have to record the context switch
801 DumpVeryRawGranEvent(TimeOfLastYield, CURRENT_PROC, CURRENT_PROC,
802 GR_DESCHEDULE, LastTSO, (StgClosure *)NULL, 0, 0);
805 if (RtsFlags.ParFlags.ParStats.Full &&
806 (emitSchedule /* forced emit */ ||
807 (t && LastTSO && t->id != LastTSO->id))) {
809 we are running a different TSO, so write a schedule event to log file
810 NB: If we use fair scheduling we also have to write a deschedule
811 event for LastTSO; with unfair scheduling we know that the
812 previous tso has blocked whenever we switch to another tso, so
813 we don't need it in GUM for now
815 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
816 GR_SCHEDULE, t, (StgClosure *)NULL, 0, 0);
817 emitSchedule = rtsFalse;
821 #else /* !GRAN && !PAR */
823 // grab a thread from the run queue
824 ASSERT(run_queue_hd != END_TSO_QUEUE);
827 // Sanity check the thread we're about to run. This can be
828 // expensive if there is lots of thread switching going on...
829 IF_DEBUG(sanity,checkTSO(t));
834 StgMainThread *m = t->main;
841 sched_belch("### Running thread %d in bound thread", t->id));
842 // yes, the Haskell thread is bound to the current native thread
847 sched_belch("### thread %d bound to another OS thread", t->id));
848 // no, bound to a different Haskell thread: pass to that thread
849 PUSH_ON_RUN_QUEUE(t);
850 passCapability(&m->bound_thread_cond);
856 if(mainThread != NULL)
857 // The thread we want to run is bound.
860 sched_belch("### this OS thread cannot run thread %d", t->id));
861 // no, the current native thread is bound to a different
862 // Haskell thread, so pass it to any worker thread
863 PUSH_ON_RUN_QUEUE(t);
864 passCapabilityToWorker();
871 cap->r.rCurrentTSO = t;
873 /* context switches are now initiated by the timer signal, unless
874 * the user specified "context switch as often as possible", with
877 if ((RtsFlags.ConcFlags.ctxtSwitchTicks == 0
878 && (run_queue_hd != END_TSO_QUEUE
879 || blocked_queue_hd != END_TSO_QUEUE
880 || sleeping_queue != END_TSO_QUEUE)))
887 RELEASE_LOCK(&sched_mutex);
889 IF_DEBUG(scheduler, sched_belch("-->> running thread %ld %s ...",
890 t->id, whatNext_strs[t->what_next]));
893 startHeapProfTimer();
896 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
897 /* Run the current thread
899 prev_what_next = t->what_next;
900 switch (prev_what_next) {
903 /* Thread already finished, return to scheduler. */
904 ret = ThreadFinished;
907 errno = t->saved_errno;
908 ret = StgRun((StgFunPtr) stg_returnToStackTop, &cap->r);
909 t->saved_errno = errno;
911 case ThreadInterpret:
912 ret = interpretBCO(cap);
915 barf("schedule: invalid what_next field");
917 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
919 /* Costs for the scheduler are assigned to CCS_SYSTEM */
925 ACQUIRE_LOCK(&sched_mutex);
927 #ifdef RTS_SUPPORTS_THREADS
928 IF_DEBUG(scheduler,fprintf(stderr,"sched (task %p): ", osThreadId()););
929 #elif !defined(GRAN) && !defined(PAR)
930 IF_DEBUG(scheduler,fprintf(stderr,"sched: "););
932 t = cap->r.rCurrentTSO;
935 /* HACK 675: if the last thread didn't yield, make sure to print a
936 SCHEDULE event to the log file when StgRunning the next thread, even
937 if it is the same one as before */
939 TimeOfLastYield = CURRENT_TIME;
945 IF_DEBUG(gran, DumpGranEvent(GR_DESCHEDULE, t));
946 globalGranStats.tot_heapover++;
948 globalParStats.tot_heapover++;
951 // did the task ask for a large block?
952 if (cap->r.rHpAlloc > BLOCK_SIZE_W) {
953 // if so, get one and push it on the front of the nursery.
957 blocks = (nat)BLOCK_ROUND_UP(cap->r.rHpAlloc * sizeof(W_)) / BLOCK_SIZE;
959 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped: requesting a large block (size %d)",
960 t->id, whatNext_strs[t->what_next], blocks));
962 // don't do this if it would push us over the
963 // alloc_blocks_lim limit; we'll GC first.
964 if (alloc_blocks + blocks < alloc_blocks_lim) {
966 alloc_blocks += blocks;
967 bd = allocGroup( blocks );
969 // link the new group into the list
970 bd->link = cap->r.rCurrentNursery;
971 bd->u.back = cap->r.rCurrentNursery->u.back;
972 if (cap->r.rCurrentNursery->u.back != NULL) {
973 cap->r.rCurrentNursery->u.back->link = bd;
975 ASSERT(g0s0->blocks == cap->r.rCurrentNursery &&
976 g0s0->blocks == cap->r.rNursery);
977 cap->r.rNursery = g0s0->blocks = bd;
979 cap->r.rCurrentNursery->u.back = bd;
981 // initialise it as a nursery block. We initialise the
982 // step, gen_no, and flags field of *every* sub-block in
983 // this large block, because this is easier than making
984 // sure that we always find the block head of a large
985 // block whenever we call Bdescr() (eg. evacuate() and
986 // isAlive() in the GC would both have to do this, at
990 for (x = bd; x < bd + blocks; x++) {
997 // don't forget to update the block count in g0s0.
998 g0s0->n_blocks += blocks;
999 // This assert can be a killer if the app is doing lots
1000 // of large block allocations.
1001 ASSERT(countBlocks(g0s0->blocks) == g0s0->n_blocks);
1003 // now update the nursery to point to the new block
1004 cap->r.rCurrentNursery = bd;
1006 // we might be unlucky and have another thread get on the
1007 // run queue before us and steal the large block, but in that
1008 // case the thread will just end up requesting another large
1010 PUSH_ON_RUN_QUEUE(t);
1015 /* make all the running tasks block on a condition variable,
1016 * maybe set context_switch and wait till they all pile in,
1017 * then have them wait on a GC condition variable.
1019 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped: HeapOverflow",
1020 t->id, whatNext_strs[t->what_next]));
1023 ASSERT(!is_on_queue(t,CurrentProc));
1025 /* Currently we emit a DESCHEDULE event before GC in GUM.
1026 ToDo: either add separate event to distinguish SYSTEM time from rest
1027 or just nuke this DESCHEDULE (and the following SCHEDULE) */
1028 if (0 && RtsFlags.ParFlags.ParStats.Full) {
1029 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1030 GR_DESCHEDULE, t, (StgClosure *)NULL, 0, 0);
1031 emitSchedule = rtsTrue;
1035 ready_to_gc = rtsTrue;
1036 context_switch = 1; /* stop other threads ASAP */
1037 PUSH_ON_RUN_QUEUE(t);
1038 /* actual GC is done at the end of the while loop */
1044 DumpGranEvent(GR_DESCHEDULE, t));
1045 globalGranStats.tot_stackover++;
1048 // DumpGranEvent(GR_DESCHEDULE, t);
1049 globalParStats.tot_stackover++;
1051 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped, StackOverflow",
1052 t->id, whatNext_strs[t->what_next]));
1053 /* just adjust the stack for this thread, then pop it back
1058 /* enlarge the stack */
1059 StgTSO *new_t = threadStackOverflow(t);
1061 /* This TSO has moved, so update any pointers to it from the
1062 * main thread stack. It better not be on any other queues...
1063 * (it shouldn't be).
1065 if (t->main != NULL) {
1066 t->main->tso = new_t;
1068 PUSH_ON_RUN_QUEUE(new_t);
1072 case ThreadYielding:
1075 DumpGranEvent(GR_DESCHEDULE, t));
1076 globalGranStats.tot_yields++;
1079 // DumpGranEvent(GR_DESCHEDULE, t);
1080 globalParStats.tot_yields++;
1082 /* put the thread back on the run queue. Then, if we're ready to
1083 * GC, check whether this is the last task to stop. If so, wake
1084 * up the GC thread. getThread will block during a GC until the
1088 if (t->what_next != prev_what_next) {
1089 belch("--<< thread %ld (%s) stopped to switch evaluators",
1090 t->id, whatNext_strs[t->what_next]);
1092 belch("--<< thread %ld (%s) stopped, yielding",
1093 t->id, whatNext_strs[t->what_next]);
1098 //belch("&& Doing sanity check on yielding TSO %ld.", t->id);
1100 ASSERT(t->link == END_TSO_QUEUE);
1102 // Shortcut if we're just switching evaluators: don't bother
1103 // doing stack squeezing (which can be expensive), just run the
1105 if (t->what_next != prev_what_next) {
1112 ASSERT(!is_on_queue(t,CurrentProc));
1115 //belch("&& Doing sanity check on all ThreadQueues (and their TSOs).");
1116 checkThreadQsSanity(rtsTrue));
1120 if (RtsFlags.ParFlags.doFairScheduling) {
1121 /* this does round-robin scheduling; good for concurrency */
1122 APPEND_TO_RUN_QUEUE(t);
1124 /* this does unfair scheduling; good for parallelism */
1125 PUSH_ON_RUN_QUEUE(t);
1128 // this does round-robin scheduling; good for concurrency
1129 APPEND_TO_RUN_QUEUE(t);
1133 /* add a ContinueThread event to actually process the thread */
1134 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1136 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1138 belch("GRAN: eventq and runnableq after adding yielded thread to queue again:");
1147 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: ",
1148 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)));
1149 if (t->block_info.closure!=(StgClosure*)NULL) print_bq(t->block_info.closure));
1151 // ??? needed; should emit block before
1153 DumpGranEvent(GR_DESCHEDULE, t));
1154 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1157 ASSERT(procStatus[CurrentProc]==Busy ||
1158 ((procStatus[CurrentProc]==Fetching) &&
1159 (t->block_info.closure!=(StgClosure*)NULL)));
1160 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1161 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1162 procStatus[CurrentProc]==Fetching))
1163 procStatus[CurrentProc] = Idle;
1167 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p with BQ: ",
1168 t->id, t, whatNext_strs[t->what_next], t->block_info.closure));
1171 if (t->block_info.closure!=(StgClosure*)NULL)
1172 print_bq(t->block_info.closure));
1174 /* Send a fetch (if BlockedOnGA) and dump event to log file */
1177 /* whatever we schedule next, we must log that schedule */
1178 emitSchedule = rtsTrue;
1181 /* don't need to do anything. Either the thread is blocked on
1182 * I/O, in which case we'll have called addToBlockedQueue
1183 * previously, or it's blocked on an MVar or Blackhole, in which
1184 * case it'll be on the relevant queue already.
1187 fprintf(stderr, "--<< thread %d (%s) stopped: ",
1188 t->id, whatNext_strs[t->what_next]);
1189 printThreadBlockage(t);
1190 fprintf(stderr, "\n"));
1193 /* Only for dumping event to log file
1194 ToDo: do I need this in GranSim, too?
1201 case ThreadFinished:
1202 /* Need to check whether this was a main thread, and if so, signal
1203 * the task that started it with the return value. If we have no
1204 * more main threads, we probably need to stop all the tasks until
1207 /* We also end up here if the thread kills itself with an
1208 * uncaught exception, see Exception.hc.
1210 IF_DEBUG(scheduler,belch("--++ thread %d (%s) finished",
1211 t->id, whatNext_strs[t->what_next]));
1213 endThread(t, CurrentProc); // clean-up the thread
1215 /* For now all are advisory -- HWL */
1216 //if(t->priority==AdvisoryPriority) ??
1217 advisory_thread_count--;
1220 if(t->dist.priority==RevalPriority)
1224 if (RtsFlags.ParFlags.ParStats.Full &&
1225 !RtsFlags.ParFlags.ParStats.Suppressed)
1226 DumpEndEvent(CURRENT_PROC, t, rtsFalse /* not mandatory */);
1230 // Check whether the thread that just completed was a main
1231 // thread, and if so return with the result.
1233 // There is an assumption here that all thread completion goes
1234 // through this point; we need to make sure that if a thread
1235 // ends up in the ThreadKilled state, that it stays on the run
1236 // queue so it can be dealt with here.
1239 #if defined(RTS_SUPPORTS_THREADS)
1242 mainThread->tso == t
1246 // We are a bound thread: this must be our thread that just
1248 ASSERT(mainThread->tso == t);
1250 if (t->what_next == ThreadComplete) {
1251 if (mainThread->ret) {
1252 // NOTE: return val is tso->sp[1] (see StgStartup.hc)
1253 *(mainThread->ret) = (StgClosure *)mainThread->tso->sp[1];
1255 mainThread->stat = Success;
1257 if (mainThread->ret) {
1258 *(mainThread->ret) = NULL;
1260 if (was_interrupted) {
1261 mainThread->stat = Interrupted;
1263 mainThread->stat = Killed;
1267 removeThreadLabel((StgWord)mainThread->tso->id);
1269 if (mainThread->prev == NULL) {
1270 main_threads = mainThread->link;
1272 mainThread->prev->link = mainThread->link;
1274 if (mainThread->link != NULL) {
1275 mainThread->link->prev = NULL;
1277 releaseCapability(cap);
1281 #ifdef RTS_SUPPORTS_THREADS
1282 ASSERT(t->main == NULL);
1284 if (t->main != NULL) {
1285 // Must be a main thread that is not the topmost one. Leave
1286 // it on the run queue until the stack has unwound to the
1287 // point where we can deal with this. Leaving it on the run
1288 // queue also ensures that the garbage collector knows about
1289 // this thread and its return value (it gets dropped from the
1290 // all_threads list so there's no other way to find it).
1291 APPEND_TO_RUN_QUEUE(t);
1297 barf("schedule: invalid thread return code %d", (int)ret);
1301 // When we have +RTS -i0 and we're heap profiling, do a census at
1302 // every GC. This lets us get repeatable runs for debugging.
1303 if (performHeapProfile ||
1304 (RtsFlags.ProfFlags.profileInterval==0 &&
1305 RtsFlags.ProfFlags.doHeapProfile && ready_to_gc)) {
1306 GarbageCollect(GetRoots, rtsTrue);
1308 performHeapProfile = rtsFalse;
1309 ready_to_gc = rtsFalse; // we already GC'd
1314 /* everybody back, start the GC.
1315 * Could do it in this thread, or signal a condition var
1316 * to do it in another thread. Either way, we need to
1317 * broadcast on gc_pending_cond afterward.
1319 #if defined(RTS_SUPPORTS_THREADS)
1320 IF_DEBUG(scheduler,sched_belch("doing GC"));
1322 GarbageCollect(GetRoots,rtsFalse);
1323 ready_to_gc = rtsFalse;
1325 /* add a ContinueThread event to continue execution of current thread */
1326 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1328 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1330 fprintf(stderr, "GRAN: eventq and runnableq after Garbage collection:\n");
1338 IF_GRAN_DEBUG(unused,
1339 print_eventq(EventHd));
1341 event = get_next_event();
1344 /* ToDo: wait for next message to arrive rather than busy wait */
1347 } /* end of while(1) */
1349 IF_PAR_DEBUG(verbose,
1350 belch("== Leaving schedule() after having received Finish"));
1353 /* ---------------------------------------------------------------------------
1354 * rtsSupportsBoundThreads(): is the RTS built to support bound threads?
1355 * used by Control.Concurrent for error checking.
1356 * ------------------------------------------------------------------------- */
1359 rtsSupportsBoundThreads(void)
1368 /* ---------------------------------------------------------------------------
1369 * isThreadBound(tso): check whether tso is bound to an OS thread.
1370 * ------------------------------------------------------------------------- */
1373 isThreadBound(StgTSO* tso USED_IN_THREADED_RTS)
1376 return (tso->main != NULL);
1381 /* ---------------------------------------------------------------------------
1382 * Singleton fork(). Do not copy any running threads.
1383 * ------------------------------------------------------------------------- */
1385 #ifndef mingw32_TARGET_OS
1386 #define FORKPROCESS_PRIMOP_SUPPORTED
1389 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1391 deleteThreadImmediately(StgTSO *tso);
1394 forkProcess(HsStablePtr *entry
1395 #ifndef FORKPROCESS_PRIMOP_SUPPORTED
1400 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1406 IF_DEBUG(scheduler,sched_belch("forking!"));
1407 rts_lock(); // This not only acquires sched_mutex, it also
1408 // makes sure that no other threads are running
1412 if (pid) { /* parent */
1414 /* just return the pid */
1418 } else { /* child */
1421 // delete all threads
1422 run_queue_hd = run_queue_tl = END_TSO_QUEUE;
1424 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1427 // don't allow threads to catch the ThreadKilled exception
1428 deleteThreadImmediately(t);
1431 // wipe the main thread list
1432 while((m = main_threads) != NULL) {
1433 main_threads = m->link;
1434 # ifdef THREADED_RTS
1435 closeCondition(&m->bound_thread_cond);
1440 # ifdef RTS_SUPPORTS_THREADS
1441 resetTaskManagerAfterFork(); // tell startTask() and friends that
1442 startingWorkerThread = rtsFalse; // we have no worker threads any more
1443 resetWorkerWakeupPipeAfterFork();
1446 rc = rts_evalStableIO(entry, NULL); // run the action
1447 rts_checkSchedStatus("forkProcess",rc);
1451 hs_exit(); // clean up and exit
1454 #else /* !FORKPROCESS_PRIMOP_SUPPORTED */
1455 barf("forkProcess#: primop not supported, sorry!\n");
1460 /* ---------------------------------------------------------------------------
1461 * deleteAllThreads(): kill all the live threads.
1463 * This is used when we catch a user interrupt (^C), before performing
1464 * any necessary cleanups and running finalizers.
1466 * Locks: sched_mutex held.
1467 * ------------------------------------------------------------------------- */
1470 deleteAllThreads ( void )
1473 IF_DEBUG(scheduler,sched_belch("deleting all threads"));
1474 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1475 next = t->global_link;
1479 // The run queue now contains a bunch of ThreadKilled threads. We
1480 // must not throw these away: the main thread(s) will be in there
1481 // somewhere, and the main scheduler loop has to deal with it.
1482 // Also, the run queue is the only thing keeping these threads from
1483 // being GC'd, and we don't want the "main thread has been GC'd" panic.
1485 ASSERT(blocked_queue_hd == END_TSO_QUEUE);
1486 ASSERT(sleeping_queue == END_TSO_QUEUE);
1489 /* startThread and insertThread are now in GranSim.c -- HWL */
1492 /* ---------------------------------------------------------------------------
1493 * Suspending & resuming Haskell threads.
1495 * When making a "safe" call to C (aka _ccall_GC), the task gives back
1496 * its capability before calling the C function. This allows another
1497 * task to pick up the capability and carry on running Haskell
1498 * threads. It also means that if the C call blocks, it won't lock
1501 * The Haskell thread making the C call is put to sleep for the
1502 * duration of the call, on the susepended_ccalling_threads queue. We
1503 * give out a token to the task, which it can use to resume the thread
1504 * on return from the C function.
1505 * ------------------------------------------------------------------------- */
1508 suspendThread( StgRegTable *reg,
1517 int saved_errno = errno;
1519 /* assume that *reg is a pointer to the StgRegTable part
1522 cap = (Capability *)((void *)((unsigned char*)reg - sizeof(StgFunTable)));
1524 ACQUIRE_LOCK(&sched_mutex);
1527 sched_belch("thread %d did a _ccall_gc (is_concurrent: %d)", cap->r.rCurrentTSO->id,concCall));
1529 // XXX this might not be necessary --SDM
1530 cap->r.rCurrentTSO->what_next = ThreadRunGHC;
1532 threadPaused(cap->r.rCurrentTSO);
1533 cap->r.rCurrentTSO->link = suspended_ccalling_threads;
1534 suspended_ccalling_threads = cap->r.rCurrentTSO;
1536 if(cap->r.rCurrentTSO->blocked_exceptions == NULL) {
1537 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall;
1538 cap->r.rCurrentTSO->blocked_exceptions = END_TSO_QUEUE;
1540 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall_NoUnblockExc;
1543 /* Use the thread ID as the token; it should be unique */
1544 tok = cap->r.rCurrentTSO->id;
1546 /* Hand back capability */
1547 releaseCapability(cap);
1549 #if defined(RTS_SUPPORTS_THREADS)
1550 /* Preparing to leave the RTS, so ensure there's a native thread/task
1551 waiting to take over.
1553 IF_DEBUG(scheduler, sched_belch("worker (token %d): leaving RTS", tok));
1556 /* Other threads _might_ be available for execution; signal this */
1558 RELEASE_LOCK(&sched_mutex);
1560 errno = saved_errno;
1565 resumeThread( StgInt tok,
1566 rtsBool concCall STG_UNUSED )
1568 StgTSO *tso, **prev;
1570 int saved_errno = errno;
1572 #if defined(RTS_SUPPORTS_THREADS)
1573 /* Wait for permission to re-enter the RTS with the result. */
1574 ACQUIRE_LOCK(&sched_mutex);
1575 waitForReturnCapability(&sched_mutex, &cap);
1577 IF_DEBUG(scheduler, sched_belch("worker (token %d): re-entering RTS", tok));
1579 grabCapability(&cap);
1582 /* Remove the thread off of the suspended list */
1583 prev = &suspended_ccalling_threads;
1584 for (tso = suspended_ccalling_threads;
1585 tso != END_TSO_QUEUE;
1586 prev = &tso->link, tso = tso->link) {
1587 if (tso->id == (StgThreadID)tok) {
1592 if (tso == END_TSO_QUEUE) {
1593 barf("resumeThread: thread not found");
1595 tso->link = END_TSO_QUEUE;
1597 if(tso->why_blocked == BlockedOnCCall) {
1598 awakenBlockedQueueNoLock(tso->blocked_exceptions);
1599 tso->blocked_exceptions = NULL;
1602 /* Reset blocking status */
1603 tso->why_blocked = NotBlocked;
1605 cap->r.rCurrentTSO = tso;
1606 RELEASE_LOCK(&sched_mutex);
1607 errno = saved_errno;
1612 /* ---------------------------------------------------------------------------
1614 * ------------------------------------------------------------------------ */
1615 static void unblockThread(StgTSO *tso);
1617 /* ---------------------------------------------------------------------------
1618 * Comparing Thread ids.
1620 * This is used from STG land in the implementation of the
1621 * instances of Eq/Ord for ThreadIds.
1622 * ------------------------------------------------------------------------ */
1625 cmp_thread(StgPtr tso1, StgPtr tso2)
1627 StgThreadID id1 = ((StgTSO *)tso1)->id;
1628 StgThreadID id2 = ((StgTSO *)tso2)->id;
1630 if (id1 < id2) return (-1);
1631 if (id1 > id2) return 1;
1635 /* ---------------------------------------------------------------------------
1636 * Fetching the ThreadID from an StgTSO.
1638 * This is used in the implementation of Show for ThreadIds.
1639 * ------------------------------------------------------------------------ */
1641 rts_getThreadId(StgPtr tso)
1643 return ((StgTSO *)tso)->id;
1648 labelThread(StgPtr tso, char *label)
1653 /* Caveat: Once set, you can only set the thread name to "" */
1654 len = strlen(label)+1;
1655 buf = stgMallocBytes(len * sizeof(char), "Schedule.c:labelThread()");
1656 strncpy(buf,label,len);
1657 /* Update will free the old memory for us */
1658 updateThreadLabel(((StgTSO *)tso)->id,buf);
1662 /* ---------------------------------------------------------------------------
1663 Create a new thread.
1665 The new thread starts with the given stack size. Before the
1666 scheduler can run, however, this thread needs to have a closure
1667 (and possibly some arguments) pushed on its stack. See
1668 pushClosure() in Schedule.h.
1670 createGenThread() and createIOThread() (in SchedAPI.h) are
1671 convenient packaged versions of this function.
1673 currently pri (priority) is only used in a GRAN setup -- HWL
1674 ------------------------------------------------------------------------ */
1676 /* currently pri (priority) is only used in a GRAN setup -- HWL */
1678 createThread(nat size, StgInt pri)
1681 createThread(nat size)
1688 /* First check whether we should create a thread at all */
1690 /* check that no more than RtsFlags.ParFlags.maxThreads threads are created */
1691 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads) {
1693 belch("{createThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1694 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1695 return END_TSO_QUEUE;
1701 ASSERT(!RtsFlags.GranFlags.Light || CurrentProc==0);
1704 // ToDo: check whether size = stack_size - TSO_STRUCT_SIZEW
1706 /* catch ridiculously small stack sizes */
1707 if (size < MIN_STACK_WORDS + TSO_STRUCT_SIZEW) {
1708 size = MIN_STACK_WORDS + TSO_STRUCT_SIZEW;
1711 stack_size = size - TSO_STRUCT_SIZEW;
1713 tso = (StgTSO *)allocate(size);
1714 TICK_ALLOC_TSO(stack_size, 0);
1716 SET_HDR(tso, &stg_TSO_info, CCS_SYSTEM);
1718 SET_GRAN_HDR(tso, ThisPE);
1721 // Always start with the compiled code evaluator
1722 tso->what_next = ThreadRunGHC;
1724 tso->id = next_thread_id++;
1725 tso->why_blocked = NotBlocked;
1726 tso->blocked_exceptions = NULL;
1728 tso->saved_errno = 0;
1731 tso->stack_size = stack_size;
1732 tso->max_stack_size = round_to_mblocks(RtsFlags.GcFlags.maxStkSize)
1734 tso->sp = (P_)&(tso->stack) + stack_size;
1737 tso->prof.CCCS = CCS_MAIN;
1740 /* put a stop frame on the stack */
1741 tso->sp -= sizeofW(StgStopFrame);
1742 SET_HDR((StgClosure*)tso->sp,(StgInfoTable *)&stg_stop_thread_info,CCS_SYSTEM);
1745 tso->link = END_TSO_QUEUE;
1746 /* uses more flexible routine in GranSim */
1747 insertThread(tso, CurrentProc);
1749 /* In a non-GranSim setup the pushing of a TSO onto the runq is separated
1755 if (RtsFlags.GranFlags.GranSimStats.Full)
1756 DumpGranEvent(GR_START,tso);
1758 if (RtsFlags.ParFlags.ParStats.Full)
1759 DumpGranEvent(GR_STARTQ,tso);
1760 /* HACk to avoid SCHEDULE
1764 /* Link the new thread on the global thread list.
1766 tso->global_link = all_threads;
1770 tso->dist.priority = MandatoryPriority; //by default that is...
1774 tso->gran.pri = pri;
1776 tso->gran.magic = TSO_MAGIC; // debugging only
1778 tso->gran.sparkname = 0;
1779 tso->gran.startedat = CURRENT_TIME;
1780 tso->gran.exported = 0;
1781 tso->gran.basicblocks = 0;
1782 tso->gran.allocs = 0;
1783 tso->gran.exectime = 0;
1784 tso->gran.fetchtime = 0;
1785 tso->gran.fetchcount = 0;
1786 tso->gran.blocktime = 0;
1787 tso->gran.blockcount = 0;
1788 tso->gran.blockedat = 0;
1789 tso->gran.globalsparks = 0;
1790 tso->gran.localsparks = 0;
1791 if (RtsFlags.GranFlags.Light)
1792 tso->gran.clock = Now; /* local clock */
1794 tso->gran.clock = 0;
1796 IF_DEBUG(gran,printTSO(tso));
1799 tso->par.magic = TSO_MAGIC; // debugging only
1801 tso->par.sparkname = 0;
1802 tso->par.startedat = CURRENT_TIME;
1803 tso->par.exported = 0;
1804 tso->par.basicblocks = 0;
1805 tso->par.allocs = 0;
1806 tso->par.exectime = 0;
1807 tso->par.fetchtime = 0;
1808 tso->par.fetchcount = 0;
1809 tso->par.blocktime = 0;
1810 tso->par.blockcount = 0;
1811 tso->par.blockedat = 0;
1812 tso->par.globalsparks = 0;
1813 tso->par.localsparks = 0;
1817 globalGranStats.tot_threads_created++;
1818 globalGranStats.threads_created_on_PE[CurrentProc]++;
1819 globalGranStats.tot_sq_len += spark_queue_len(CurrentProc);
1820 globalGranStats.tot_sq_probes++;
1822 // collect parallel global statistics (currently done together with GC stats)
1823 if (RtsFlags.ParFlags.ParStats.Global &&
1824 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
1825 //fprintf(stderr, "Creating thread %d @ %11.2f\n", tso->id, usertime());
1826 globalParStats.tot_threads_created++;
1832 belch("==__ schedule: Created TSO %d (%p);",
1833 CurrentProc, tso, tso->id));
1835 IF_PAR_DEBUG(verbose,
1836 belch("==__ schedule: Created TSO %d (%p); %d threads active",
1837 tso->id, tso, advisory_thread_count));
1839 IF_DEBUG(scheduler,sched_belch("created thread %ld, stack size = %lx words",
1840 tso->id, tso->stack_size));
1847 all parallel thread creation calls should fall through the following routine.
1850 createSparkThread(rtsSpark spark)
1852 ASSERT(spark != (rtsSpark)NULL);
1853 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads)
1855 barf("{createSparkThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1856 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1857 return END_TSO_QUEUE;
1861 tso = createThread(RtsFlags.GcFlags.initialStkSize);
1862 if (tso==END_TSO_QUEUE)
1863 barf("createSparkThread: Cannot create TSO");
1865 tso->priority = AdvisoryPriority;
1867 pushClosure(tso,spark);
1868 PUSH_ON_RUN_QUEUE(tso);
1869 advisory_thread_count++;
1876 Turn a spark into a thread.
1877 ToDo: fix for SMP (needs to acquire SCHED_MUTEX!)
1881 activateSpark (rtsSpark spark)
1885 tso = createSparkThread(spark);
1886 if (RtsFlags.ParFlags.ParStats.Full) {
1887 //ASSERT(run_queue_hd == END_TSO_QUEUE); // I think ...
1888 IF_PAR_DEBUG(verbose,
1889 belch("==^^ activateSpark: turning spark of closure %p (%s) into a thread",
1890 (StgClosure *)spark, info_type((StgClosure *)spark)));
1892 // ToDo: fwd info on local/global spark to thread -- HWL
1893 // tso->gran.exported = spark->exported;
1894 // tso->gran.locked = !spark->global;
1895 // tso->gran.sparkname = spark->name;
1901 static SchedulerStatus waitThread_(/*out*/StgMainThread* m,
1902 Capability *initialCapability
1906 /* ---------------------------------------------------------------------------
1909 * scheduleThread puts a thread on the head of the runnable queue.
1910 * This will usually be done immediately after a thread is created.
1911 * The caller of scheduleThread must create the thread using e.g.
1912 * createThread and push an appropriate closure
1913 * on this thread's stack before the scheduler is invoked.
1914 * ------------------------------------------------------------------------ */
1916 static void scheduleThread_ (StgTSO* tso);
1919 scheduleThread_(StgTSO *tso)
1921 // Precondition: sched_mutex must be held.
1922 PUSH_ON_RUN_QUEUE(tso);
1927 scheduleThread(StgTSO* tso)
1929 ACQUIRE_LOCK(&sched_mutex);
1930 scheduleThread_(tso);
1931 RELEASE_LOCK(&sched_mutex);
1934 #if defined(RTS_SUPPORTS_THREADS)
1935 static Condition bound_cond_cache;
1936 static int bound_cond_cache_full = 0;
1941 scheduleWaitThread(StgTSO* tso, /*[out]*/HaskellObj* ret,
1942 Capability *initialCapability)
1944 // Precondition: sched_mutex must be held
1947 m = stgMallocBytes(sizeof(StgMainThread), "waitThread");
1952 m->link = main_threads;
1954 if (main_threads != NULL) {
1955 main_threads->prev = m;
1959 #if defined(RTS_SUPPORTS_THREADS)
1960 // Allocating a new condition for each thread is expensive, so we
1961 // cache one. This is a pretty feeble hack, but it helps speed up
1962 // consecutive call-ins quite a bit.
1963 if (bound_cond_cache_full) {
1964 m->bound_thread_cond = bound_cond_cache;
1965 bound_cond_cache_full = 0;
1967 initCondition(&m->bound_thread_cond);
1971 /* Put the thread on the main-threads list prior to scheduling the TSO.
1972 Failure to do so introduces a race condition in the MT case (as
1973 identified by Wolfgang Thaller), whereby the new task/OS thread
1974 created by scheduleThread_() would complete prior to the thread
1975 that spawned it managed to put 'itself' on the main-threads list.
1976 The upshot of it all being that the worker thread wouldn't get to
1977 signal the completion of the its work item for the main thread to
1978 see (==> it got stuck waiting.) -- sof 6/02.
1980 IF_DEBUG(scheduler, sched_belch("waiting for thread (%d)", tso->id));
1982 PUSH_ON_RUN_QUEUE(tso);
1983 // NB. Don't call THREAD_RUNNABLE() here, because the thread is
1984 // bound and only runnable by *this* OS thread, so waking up other
1985 // workers will just slow things down.
1987 return waitThread_(m, initialCapability);
1990 /* ---------------------------------------------------------------------------
1993 * Initialise the scheduler. This resets all the queues - if the
1994 * queues contained any threads, they'll be garbage collected at the
1997 * ------------------------------------------------------------------------ */
2005 for (i=0; i<=MAX_PROC; i++) {
2006 run_queue_hds[i] = END_TSO_QUEUE;
2007 run_queue_tls[i] = END_TSO_QUEUE;
2008 blocked_queue_hds[i] = END_TSO_QUEUE;
2009 blocked_queue_tls[i] = END_TSO_QUEUE;
2010 ccalling_threadss[i] = END_TSO_QUEUE;
2011 sleeping_queue = END_TSO_QUEUE;
2014 run_queue_hd = END_TSO_QUEUE;
2015 run_queue_tl = END_TSO_QUEUE;
2016 blocked_queue_hd = END_TSO_QUEUE;
2017 blocked_queue_tl = END_TSO_QUEUE;
2018 sleeping_queue = END_TSO_QUEUE;
2021 suspended_ccalling_threads = END_TSO_QUEUE;
2023 main_threads = NULL;
2024 all_threads = END_TSO_QUEUE;
2029 RtsFlags.ConcFlags.ctxtSwitchTicks =
2030 RtsFlags.ConcFlags.ctxtSwitchTime / TICK_MILLISECS;
2032 #if defined(RTS_SUPPORTS_THREADS)
2033 /* Initialise the mutex and condition variables used by
2035 initMutex(&sched_mutex);
2036 initMutex(&term_mutex);
2039 ACQUIRE_LOCK(&sched_mutex);
2041 /* A capability holds the state a native thread needs in
2042 * order to execute STG code. At least one capability is
2043 * floating around (only SMP builds have more than one).
2047 #if defined(RTS_SUPPORTS_THREADS)
2048 /* start our haskell execution tasks */
2049 startTaskManager(0,taskStart);
2052 #if /* defined(SMP) ||*/ defined(PAR)
2056 RELEASE_LOCK(&sched_mutex);
2060 exitScheduler( void )
2062 #if defined(RTS_SUPPORTS_THREADS)
2065 shutting_down_scheduler = rtsTrue;
2068 /* ----------------------------------------------------------------------------
2069 Managing the per-task allocation areas.
2071 Each capability comes with an allocation area. These are
2072 fixed-length block lists into which allocation can be done.
2074 ToDo: no support for two-space collection at the moment???
2075 ------------------------------------------------------------------------- */
2079 waitThread_(StgMainThread* m, Capability *initialCapability)
2081 SchedulerStatus stat;
2083 // Precondition: sched_mutex must be held.
2084 IF_DEBUG(scheduler, sched_belch("new main thread (%d)", m->tso->id));
2087 /* GranSim specific init */
2088 CurrentTSO = m->tso; // the TSO to run
2089 procStatus[MainProc] = Busy; // status of main PE
2090 CurrentProc = MainProc; // PE to run it on
2091 schedule(m,initialCapability);
2093 schedule(m,initialCapability);
2094 ASSERT(m->stat != NoStatus);
2099 #if defined(RTS_SUPPORTS_THREADS)
2100 // Free the condition variable, returning it to the cache if possible.
2101 if (!bound_cond_cache_full) {
2102 bound_cond_cache = m->bound_thread_cond;
2103 bound_cond_cache_full = 1;
2105 closeCondition(&m->bound_thread_cond);
2109 IF_DEBUG(scheduler, sched_belch("main thread (%d) finished", m->tso->id));
2112 // Postcondition: sched_mutex still held
2116 /* ---------------------------------------------------------------------------
2117 Where are the roots that we know about?
2119 - all the threads on the runnable queue
2120 - all the threads on the blocked queue
2121 - all the threads on the sleeping queue
2122 - all the thread currently executing a _ccall_GC
2123 - all the "main threads"
2125 ------------------------------------------------------------------------ */
2127 /* This has to be protected either by the scheduler monitor, or by the
2128 garbage collection monitor (probably the latter).
2133 GetRoots( evac_fn evac )
2138 for (i=0; i<=RtsFlags.GranFlags.proc; i++) {
2139 if ((run_queue_hds[i] != END_TSO_QUEUE) && ((run_queue_hds[i] != NULL)))
2140 evac((StgClosure **)&run_queue_hds[i]);
2141 if ((run_queue_tls[i] != END_TSO_QUEUE) && ((run_queue_tls[i] != NULL)))
2142 evac((StgClosure **)&run_queue_tls[i]);
2144 if ((blocked_queue_hds[i] != END_TSO_QUEUE) && ((blocked_queue_hds[i] != NULL)))
2145 evac((StgClosure **)&blocked_queue_hds[i]);
2146 if ((blocked_queue_tls[i] != END_TSO_QUEUE) && ((blocked_queue_tls[i] != NULL)))
2147 evac((StgClosure **)&blocked_queue_tls[i]);
2148 if ((ccalling_threadss[i] != END_TSO_QUEUE) && ((ccalling_threadss[i] != NULL)))
2149 evac((StgClosure **)&ccalling_threads[i]);
2156 if (run_queue_hd != END_TSO_QUEUE) {
2157 ASSERT(run_queue_tl != END_TSO_QUEUE);
2158 evac((StgClosure **)&run_queue_hd);
2159 evac((StgClosure **)&run_queue_tl);
2162 if (blocked_queue_hd != END_TSO_QUEUE) {
2163 ASSERT(blocked_queue_tl != END_TSO_QUEUE);
2164 evac((StgClosure **)&blocked_queue_hd);
2165 evac((StgClosure **)&blocked_queue_tl);
2168 if (sleeping_queue != END_TSO_QUEUE) {
2169 evac((StgClosure **)&sleeping_queue);
2173 if (suspended_ccalling_threads != END_TSO_QUEUE) {
2174 evac((StgClosure **)&suspended_ccalling_threads);
2177 #if defined(PAR) || defined(GRAN)
2178 markSparkQueue(evac);
2181 #if defined(RTS_USER_SIGNALS)
2182 // mark the signal handlers (signals should be already blocked)
2183 markSignalHandlers(evac);
2187 /* -----------------------------------------------------------------------------
2190 This is the interface to the garbage collector from Haskell land.
2191 We provide this so that external C code can allocate and garbage
2192 collect when called from Haskell via _ccall_GC.
2194 It might be useful to provide an interface whereby the programmer
2195 can specify more roots (ToDo).
2197 This needs to be protected by the GC condition variable above. KH.
2198 -------------------------------------------------------------------------- */
2200 static void (*extra_roots)(evac_fn);
2205 /* Obligated to hold this lock upon entry */
2206 ACQUIRE_LOCK(&sched_mutex);
2207 GarbageCollect(GetRoots,rtsFalse);
2208 RELEASE_LOCK(&sched_mutex);
2212 performMajorGC(void)
2214 ACQUIRE_LOCK(&sched_mutex);
2215 GarbageCollect(GetRoots,rtsTrue);
2216 RELEASE_LOCK(&sched_mutex);
2220 AllRoots(evac_fn evac)
2222 GetRoots(evac); // the scheduler's roots
2223 extra_roots(evac); // the user's roots
2227 performGCWithRoots(void (*get_roots)(evac_fn))
2229 ACQUIRE_LOCK(&sched_mutex);
2230 extra_roots = get_roots;
2231 GarbageCollect(AllRoots,rtsFalse);
2232 RELEASE_LOCK(&sched_mutex);
2235 /* -----------------------------------------------------------------------------
2238 If the thread has reached its maximum stack size, then raise the
2239 StackOverflow exception in the offending thread. Otherwise
2240 relocate the TSO into a larger chunk of memory and adjust its stack
2242 -------------------------------------------------------------------------- */
2245 threadStackOverflow(StgTSO *tso)
2247 nat new_stack_size, new_tso_size, stack_words;
2251 IF_DEBUG(sanity,checkTSO(tso));
2252 if (tso->stack_size >= tso->max_stack_size) {
2255 belch("@@ threadStackOverflow of TSO %d (%p): stack too large (now %ld; max is %ld)",
2256 tso->id, tso, tso->stack_size, tso->max_stack_size);
2257 /* If we're debugging, just print out the top of the stack */
2258 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2261 /* Send this thread the StackOverflow exception */
2262 raiseAsync(tso, (StgClosure *)stackOverflow_closure);
2266 /* Try to double the current stack size. If that takes us over the
2267 * maximum stack size for this thread, then use the maximum instead.
2268 * Finally round up so the TSO ends up as a whole number of blocks.
2270 new_stack_size = stg_min(tso->stack_size * 2, tso->max_stack_size);
2271 new_tso_size = (nat)BLOCK_ROUND_UP(new_stack_size * sizeof(W_) +
2272 TSO_STRUCT_SIZE)/sizeof(W_);
2273 new_tso_size = round_to_mblocks(new_tso_size); /* Be MBLOCK-friendly */
2274 new_stack_size = new_tso_size - TSO_STRUCT_SIZEW;
2276 IF_DEBUG(scheduler, fprintf(stderr,"== sched: increasing stack size from %d words to %d.\n", tso->stack_size, new_stack_size));
2278 dest = (StgTSO *)allocate(new_tso_size);
2279 TICK_ALLOC_TSO(new_stack_size,0);
2281 /* copy the TSO block and the old stack into the new area */
2282 memcpy(dest,tso,TSO_STRUCT_SIZE);
2283 stack_words = tso->stack + tso->stack_size - tso->sp;
2284 new_sp = (P_)dest + new_tso_size - stack_words;
2285 memcpy(new_sp, tso->sp, stack_words * sizeof(W_));
2287 /* relocate the stack pointers... */
2289 dest->stack_size = new_stack_size;
2291 /* Mark the old TSO as relocated. We have to check for relocated
2292 * TSOs in the garbage collector and any primops that deal with TSOs.
2294 * It's important to set the sp value to just beyond the end
2295 * of the stack, so we don't attempt to scavenge any part of the
2298 tso->what_next = ThreadRelocated;
2300 tso->sp = (P_)&(tso->stack[tso->stack_size]);
2301 tso->why_blocked = NotBlocked;
2302 dest->mut_link = NULL;
2304 IF_PAR_DEBUG(verbose,
2305 belch("@@ threadStackOverflow of TSO %d (now at %p): stack size increased to %ld",
2306 tso->id, tso, tso->stack_size);
2307 /* If we're debugging, just print out the top of the stack */
2308 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2311 IF_DEBUG(sanity,checkTSO(tso));
2313 IF_DEBUG(scheduler,printTSO(dest));
2319 /* ---------------------------------------------------------------------------
2320 Wake up a queue that was blocked on some resource.
2321 ------------------------------------------------------------------------ */
2325 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2330 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2332 /* write RESUME events to log file and
2333 update blocked and fetch time (depending on type of the orig closure) */
2334 if (RtsFlags.ParFlags.ParStats.Full) {
2335 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
2336 GR_RESUMEQ, ((StgTSO *)bqe), ((StgTSO *)bqe)->block_info.closure,
2337 0, 0 /* spark_queue_len(ADVISORY_POOL) */);
2338 if (EMPTY_RUN_QUEUE())
2339 emitSchedule = rtsTrue;
2341 switch (get_itbl(node)->type) {
2343 ((StgTSO *)bqe)->par.fetchtime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2348 ((StgTSO *)bqe)->par.blocktime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2355 barf("{unblockOneLocked}Daq Qagh: unexpected closure in blocking queue");
2362 static StgBlockingQueueElement *
2363 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2366 PEs node_loc, tso_loc;
2368 node_loc = where_is(node); // should be lifted out of loop
2369 tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2370 tso_loc = where_is((StgClosure *)tso);
2371 if (IS_LOCAL_TO(PROCS(node),tso_loc)) { // TSO is local
2372 /* !fake_fetch => TSO is on CurrentProc is same as IS_LOCAL_TO */
2373 ASSERT(CurrentProc!=node_loc || tso_loc==CurrentProc);
2374 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.lunblocktime;
2375 // insertThread(tso, node_loc);
2376 new_event(tso_loc, tso_loc, CurrentTime[CurrentProc],
2378 tso, node, (rtsSpark*)NULL);
2379 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2382 } else { // TSO is remote (actually should be FMBQ)
2383 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.mpacktime +
2384 RtsFlags.GranFlags.Costs.gunblocktime +
2385 RtsFlags.GranFlags.Costs.latency;
2386 new_event(tso_loc, CurrentProc, CurrentTime[CurrentProc],
2388 tso, node, (rtsSpark*)NULL);
2389 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2392 /* the thread-queue-overhead is accounted for in either Resume or UnblockThread */
2394 fprintf(stderr," %s TSO %d (%p) [PE %d] (block_info.closure=%p) (next=%p) ,",
2395 (node_loc==tso_loc ? "Local" : "Global"),
2396 tso->id, tso, CurrentProc, tso->block_info.closure, tso->link));
2397 tso->block_info.closure = NULL;
2398 IF_DEBUG(scheduler,belch("-- Waking up thread %ld (%p)",
2402 static StgBlockingQueueElement *
2403 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2405 StgBlockingQueueElement *next;
2407 switch (get_itbl(bqe)->type) {
2409 ASSERT(((StgTSO *)bqe)->why_blocked != NotBlocked);
2410 /* if it's a TSO just push it onto the run_queue */
2412 // ((StgTSO *)bqe)->link = END_TSO_QUEUE; // debugging?
2413 PUSH_ON_RUN_QUEUE((StgTSO *)bqe);
2415 unblockCount(bqe, node);
2416 /* reset blocking status after dumping event */
2417 ((StgTSO *)bqe)->why_blocked = NotBlocked;
2421 /* if it's a BLOCKED_FETCH put it on the PendingFetches list */
2423 bqe->link = (StgBlockingQueueElement *)PendingFetches;
2424 PendingFetches = (StgBlockedFetch *)bqe;
2428 /* can ignore this case in a non-debugging setup;
2429 see comments on RBHSave closures above */
2431 /* check that the closure is an RBHSave closure */
2432 ASSERT(get_itbl((StgClosure *)bqe) == &stg_RBH_Save_0_info ||
2433 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_1_info ||
2434 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_2_info);
2438 barf("{unblockOneLocked}Daq Qagh: Unexpected IP (%#lx; %s) in blocking queue at %#lx\n",
2439 get_itbl((StgClosure *)bqe), info_type((StgClosure *)bqe),
2443 IF_PAR_DEBUG(bq, fprintf(stderr, ", %p (%s)", bqe, info_type((StgClosure*)bqe)));
2447 #else /* !GRAN && !PAR */
2449 unblockOneLocked(StgTSO *tso)
2453 ASSERT(get_itbl(tso)->type == TSO);
2454 ASSERT(tso->why_blocked != NotBlocked);
2455 tso->why_blocked = NotBlocked;
2457 PUSH_ON_RUN_QUEUE(tso);
2459 IF_DEBUG(scheduler,sched_belch("waking up thread %ld", tso->id));
2464 #if defined(GRAN) || defined(PAR)
2465 INLINE_ME StgBlockingQueueElement *
2466 unblockOne(StgBlockingQueueElement *bqe, StgClosure *node)
2468 ACQUIRE_LOCK(&sched_mutex);
2469 bqe = unblockOneLocked(bqe, node);
2470 RELEASE_LOCK(&sched_mutex);
2475 unblockOne(StgTSO *tso)
2477 ACQUIRE_LOCK(&sched_mutex);
2478 tso = unblockOneLocked(tso);
2479 RELEASE_LOCK(&sched_mutex);
2486 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2488 StgBlockingQueueElement *bqe;
2493 belch("##-_ AwBQ for node %p on PE %d @ %ld by TSO %d (%p): ", \
2494 node, CurrentProc, CurrentTime[CurrentProc],
2495 CurrentTSO->id, CurrentTSO));
2497 node_loc = where_is(node);
2499 ASSERT(q == END_BQ_QUEUE ||
2500 get_itbl(q)->type == TSO || // q is either a TSO or an RBHSave
2501 get_itbl(q)->type == CONSTR); // closure (type constructor)
2502 ASSERT(is_unique(node));
2504 /* FAKE FETCH: magically copy the node to the tso's proc;
2505 no Fetch necessary because in reality the node should not have been
2506 moved to the other PE in the first place
2508 if (CurrentProc!=node_loc) {
2510 belch("## node %p is on PE %d but CurrentProc is %d (TSO %d); assuming fake fetch and adjusting bitmask (old: %#x)",
2511 node, node_loc, CurrentProc, CurrentTSO->id,
2512 // CurrentTSO, where_is(CurrentTSO),
2513 node->header.gran.procs));
2514 node->header.gran.procs = (node->header.gran.procs) | PE_NUMBER(CurrentProc);
2516 belch("## new bitmask of node %p is %#x",
2517 node, node->header.gran.procs));
2518 if (RtsFlags.GranFlags.GranSimStats.Global) {
2519 globalGranStats.tot_fake_fetches++;
2524 // ToDo: check: ASSERT(CurrentProc==node_loc);
2525 while (get_itbl(bqe)->type==TSO) { // q != END_TSO_QUEUE) {
2528 bqe points to the current element in the queue
2529 next points to the next element in the queue
2531 //tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2532 //tso_loc = where_is(tso);
2534 bqe = unblockOneLocked(bqe, node);
2537 /* if this is the BQ of an RBH, we have to put back the info ripped out of
2538 the closure to make room for the anchor of the BQ */
2539 if (bqe!=END_BQ_QUEUE) {
2540 ASSERT(get_itbl(node)->type == RBH && get_itbl(bqe)->type == CONSTR);
2542 ASSERT((info_ptr==&RBH_Save_0_info) ||
2543 (info_ptr==&RBH_Save_1_info) ||
2544 (info_ptr==&RBH_Save_2_info));
2546 /* cf. convertToRBH in RBH.c for writing the RBHSave closure */
2547 ((StgRBH *)node)->blocking_queue = (StgBlockingQueueElement *)((StgRBHSave *)bqe)->payload[0];
2548 ((StgRBH *)node)->mut_link = (StgMutClosure *)((StgRBHSave *)bqe)->payload[1];
2551 belch("## Filled in RBH_Save for %p (%s) at end of AwBQ",
2552 node, info_type(node)));
2555 /* statistics gathering */
2556 if (RtsFlags.GranFlags.GranSimStats.Global) {
2557 // globalGranStats.tot_bq_processing_time += bq_processing_time;
2558 globalGranStats.tot_bq_len += len; // total length of all bqs awakened
2559 // globalGranStats.tot_bq_len_local += len_local; // same for local TSOs only
2560 globalGranStats.tot_awbq++; // total no. of bqs awakened
2563 fprintf(stderr,"## BQ Stats of %p: [%d entries] %s\n",
2564 node, len, (bqe!=END_BQ_QUEUE) ? "RBH" : ""));
2568 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2570 StgBlockingQueueElement *bqe;
2572 ACQUIRE_LOCK(&sched_mutex);
2574 IF_PAR_DEBUG(verbose,
2575 belch("##-_ AwBQ for node %p on [%x]: ",
2579 if(get_itbl(q)->type == CONSTR || q==END_BQ_QUEUE) {
2580 IF_PAR_DEBUG(verbose, belch("## ... nothing to unblock so lets just return. RFP (BUG?)"));
2585 ASSERT(q == END_BQ_QUEUE ||
2586 get_itbl(q)->type == TSO ||
2587 get_itbl(q)->type == BLOCKED_FETCH ||
2588 get_itbl(q)->type == CONSTR);
2591 while (get_itbl(bqe)->type==TSO ||
2592 get_itbl(bqe)->type==BLOCKED_FETCH) {
2593 bqe = unblockOneLocked(bqe, node);
2595 RELEASE_LOCK(&sched_mutex);
2598 #else /* !GRAN && !PAR */
2601 awakenBlockedQueueNoLock(StgTSO *tso)
2603 while (tso != END_TSO_QUEUE) {
2604 tso = unblockOneLocked(tso);
2609 awakenBlockedQueue(StgTSO *tso)
2611 ACQUIRE_LOCK(&sched_mutex);
2612 while (tso != END_TSO_QUEUE) {
2613 tso = unblockOneLocked(tso);
2615 RELEASE_LOCK(&sched_mutex);
2619 /* ---------------------------------------------------------------------------
2621 - usually called inside a signal handler so it mustn't do anything fancy.
2622 ------------------------------------------------------------------------ */
2625 interruptStgRts(void)
2629 #ifdef RTS_SUPPORTS_THREADS
2630 wakeBlockedWorkerThread();
2634 /* -----------------------------------------------------------------------------
2637 This is for use when we raise an exception in another thread, which
2639 This has nothing to do with the UnblockThread event in GranSim. -- HWL
2640 -------------------------------------------------------------------------- */
2642 #if defined(GRAN) || defined(PAR)
2644 NB: only the type of the blocking queue is different in GranSim and GUM
2645 the operations on the queue-elements are the same
2646 long live polymorphism!
2648 Locks: sched_mutex is held upon entry and exit.
2652 unblockThread(StgTSO *tso)
2654 StgBlockingQueueElement *t, **last;
2656 switch (tso->why_blocked) {
2659 return; /* not blocked */
2662 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
2664 StgBlockingQueueElement *last_tso = END_BQ_QUEUE;
2665 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
2667 last = (StgBlockingQueueElement **)&mvar->head;
2668 for (t = (StgBlockingQueueElement *)mvar->head;
2670 last = &t->link, last_tso = t, t = t->link) {
2671 if (t == (StgBlockingQueueElement *)tso) {
2672 *last = (StgBlockingQueueElement *)tso->link;
2673 if (mvar->tail == tso) {
2674 mvar->tail = (StgTSO *)last_tso;
2679 barf("unblockThread (MVAR): TSO not found");
2682 case BlockedOnBlackHole:
2683 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
2685 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
2687 last = &bq->blocking_queue;
2688 for (t = bq->blocking_queue;
2690 last = &t->link, t = t->link) {
2691 if (t == (StgBlockingQueueElement *)tso) {
2692 *last = (StgBlockingQueueElement *)tso->link;
2696 barf("unblockThread (BLACKHOLE): TSO not found");
2699 case BlockedOnException:
2701 StgTSO *target = tso->block_info.tso;
2703 ASSERT(get_itbl(target)->type == TSO);
2705 if (target->what_next == ThreadRelocated) {
2706 target = target->link;
2707 ASSERT(get_itbl(target)->type == TSO);
2710 ASSERT(target->blocked_exceptions != NULL);
2712 last = (StgBlockingQueueElement **)&target->blocked_exceptions;
2713 for (t = (StgBlockingQueueElement *)target->blocked_exceptions;
2715 last = &t->link, t = t->link) {
2716 ASSERT(get_itbl(t)->type == TSO);
2717 if (t == (StgBlockingQueueElement *)tso) {
2718 *last = (StgBlockingQueueElement *)tso->link;
2722 barf("unblockThread (Exception): TSO not found");
2726 case BlockedOnWrite:
2727 #if defined(mingw32_TARGET_OS)
2728 case BlockedOnDoProc:
2731 /* take TSO off blocked_queue */
2732 StgBlockingQueueElement *prev = NULL;
2733 for (t = (StgBlockingQueueElement *)blocked_queue_hd; t != END_BQ_QUEUE;
2734 prev = t, t = t->link) {
2735 if (t == (StgBlockingQueueElement *)tso) {
2737 blocked_queue_hd = (StgTSO *)t->link;
2738 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
2739 blocked_queue_tl = END_TSO_QUEUE;
2742 prev->link = t->link;
2743 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
2744 blocked_queue_tl = (StgTSO *)prev;
2750 barf("unblockThread (I/O): TSO not found");
2753 case BlockedOnDelay:
2755 /* take TSO off sleeping_queue */
2756 StgBlockingQueueElement *prev = NULL;
2757 for (t = (StgBlockingQueueElement *)sleeping_queue; t != END_BQ_QUEUE;
2758 prev = t, t = t->link) {
2759 if (t == (StgBlockingQueueElement *)tso) {
2761 sleeping_queue = (StgTSO *)t->link;
2763 prev->link = t->link;
2768 barf("unblockThread (delay): TSO not found");
2772 barf("unblockThread");
2776 tso->link = END_TSO_QUEUE;
2777 tso->why_blocked = NotBlocked;
2778 tso->block_info.closure = NULL;
2779 PUSH_ON_RUN_QUEUE(tso);
2783 unblockThread(StgTSO *tso)
2787 /* To avoid locking unnecessarily. */
2788 if (tso->why_blocked == NotBlocked) {
2792 switch (tso->why_blocked) {
2795 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
2797 StgTSO *last_tso = END_TSO_QUEUE;
2798 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
2801 for (t = mvar->head; t != END_TSO_QUEUE;
2802 last = &t->link, last_tso = t, t = t->link) {
2805 if (mvar->tail == tso) {
2806 mvar->tail = last_tso;
2811 barf("unblockThread (MVAR): TSO not found");
2814 case BlockedOnBlackHole:
2815 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
2817 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
2819 last = &bq->blocking_queue;
2820 for (t = bq->blocking_queue; t != END_TSO_QUEUE;
2821 last = &t->link, t = t->link) {
2827 barf("unblockThread (BLACKHOLE): TSO not found");
2830 case BlockedOnException:
2832 StgTSO *target = tso->block_info.tso;
2834 ASSERT(get_itbl(target)->type == TSO);
2836 while (target->what_next == ThreadRelocated) {
2837 target = target->link;
2838 ASSERT(get_itbl(target)->type == TSO);
2841 ASSERT(target->blocked_exceptions != NULL);
2843 last = &target->blocked_exceptions;
2844 for (t = target->blocked_exceptions; t != END_TSO_QUEUE;
2845 last = &t->link, t = t->link) {
2846 ASSERT(get_itbl(t)->type == TSO);
2852 barf("unblockThread (Exception): TSO not found");
2856 case BlockedOnWrite:
2857 #if defined(mingw32_TARGET_OS)
2858 case BlockedOnDoProc:
2861 StgTSO *prev = NULL;
2862 for (t = blocked_queue_hd; t != END_TSO_QUEUE;
2863 prev = t, t = t->link) {
2866 blocked_queue_hd = t->link;
2867 if (blocked_queue_tl == t) {
2868 blocked_queue_tl = END_TSO_QUEUE;
2871 prev->link = t->link;
2872 if (blocked_queue_tl == t) {
2873 blocked_queue_tl = prev;
2879 barf("unblockThread (I/O): TSO not found");
2882 case BlockedOnDelay:
2884 StgTSO *prev = NULL;
2885 for (t = sleeping_queue; t != END_TSO_QUEUE;
2886 prev = t, t = t->link) {
2889 sleeping_queue = t->link;
2891 prev->link = t->link;
2896 barf("unblockThread (delay): TSO not found");
2900 barf("unblockThread");
2904 tso->link = END_TSO_QUEUE;
2905 tso->why_blocked = NotBlocked;
2906 tso->block_info.closure = NULL;
2907 PUSH_ON_RUN_QUEUE(tso);
2911 /* -----------------------------------------------------------------------------
2914 * The following function implements the magic for raising an
2915 * asynchronous exception in an existing thread.
2917 * We first remove the thread from any queue on which it might be
2918 * blocked. The possible blockages are MVARs and BLACKHOLE_BQs.
2920 * We strip the stack down to the innermost CATCH_FRAME, building
2921 * thunks in the heap for all the active computations, so they can
2922 * be restarted if necessary. When we reach a CATCH_FRAME, we build
2923 * an application of the handler to the exception, and push it on
2924 * the top of the stack.
2926 * How exactly do we save all the active computations? We create an
2927 * AP_STACK for every UpdateFrame on the stack. Entering one of these
2928 * AP_STACKs pushes everything from the corresponding update frame
2929 * upwards onto the stack. (Actually, it pushes everything up to the
2930 * next update frame plus a pointer to the next AP_STACK object.
2931 * Entering the next AP_STACK object pushes more onto the stack until we
2932 * reach the last AP_STACK object - at which point the stack should look
2933 * exactly as it did when we killed the TSO and we can continue
2934 * execution by entering the closure on top of the stack.
2936 * We can also kill a thread entirely - this happens if either (a) the
2937 * exception passed to raiseAsync is NULL, or (b) there's no
2938 * CATCH_FRAME on the stack. In either case, we strip the entire
2939 * stack and replace the thread with a zombie.
2941 * Locks: sched_mutex held upon entry nor exit.
2943 * -------------------------------------------------------------------------- */
2946 deleteThread(StgTSO *tso)
2948 raiseAsync(tso,NULL);
2951 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
2953 deleteThreadImmediately(StgTSO *tso)
2954 { // for forkProcess only:
2955 // delete thread without giving it a chance to catch the KillThread exception
2957 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
2961 if (tso->why_blocked != BlockedOnCCall &&
2962 tso->why_blocked != BlockedOnCCall_NoUnblockExc) {
2966 tso->what_next = ThreadKilled;
2971 raiseAsyncWithLock(StgTSO *tso, StgClosure *exception)
2973 /* When raising async exs from contexts where sched_mutex isn't held;
2974 use raiseAsyncWithLock(). */
2975 ACQUIRE_LOCK(&sched_mutex);
2976 raiseAsync(tso,exception);
2977 RELEASE_LOCK(&sched_mutex);
2981 raiseAsync(StgTSO *tso, StgClosure *exception)
2983 StgRetInfoTable *info;
2986 // Thread already dead?
2987 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
2992 sched_belch("raising exception in thread %ld.", tso->id));
2994 // Remove it from any blocking queues
2999 // The stack freezing code assumes there's a closure pointer on
3000 // the top of the stack, so we have to arrange that this is the case...
3002 if (sp[0] == (W_)&stg_enter_info) {
3006 sp[0] = (W_)&stg_dummy_ret_closure;
3012 // 1. Let the top of the stack be the "current closure"
3014 // 2. Walk up the stack until we find either an UPDATE_FRAME or a
3017 // 3. If it's an UPDATE_FRAME, then make an AP_STACK containing the
3018 // current closure applied to the chunk of stack up to (but not
3019 // including) the update frame. This closure becomes the "current
3020 // closure". Go back to step 2.
3022 // 4. If it's a CATCH_FRAME, then leave the exception handler on
3023 // top of the stack applied to the exception.
3025 // 5. If it's a STOP_FRAME, then kill the thread.
3030 info = get_ret_itbl((StgClosure *)frame);
3032 while (info->i.type != UPDATE_FRAME
3033 && (info->i.type != CATCH_FRAME || exception == NULL)
3034 && info->i.type != STOP_FRAME) {
3035 frame += stack_frame_sizeW((StgClosure *)frame);
3036 info = get_ret_itbl((StgClosure *)frame);
3039 switch (info->i.type) {
3042 // If we find a CATCH_FRAME, and we've got an exception to raise,
3043 // then build the THUNK raise(exception), and leave it on
3044 // top of the CATCH_FRAME ready to enter.
3048 StgCatchFrame *cf = (StgCatchFrame *)frame;
3052 // we've got an exception to raise, so let's pass it to the
3053 // handler in this frame.
3055 raise = (StgClosure *)allocate(sizeofW(StgClosure)+1);
3056 TICK_ALLOC_SE_THK(1,0);
3057 SET_HDR(raise,&stg_raise_info,cf->header.prof.ccs);
3058 raise->payload[0] = exception;
3060 // throw away the stack from Sp up to the CATCH_FRAME.
3064 /* Ensure that async excpetions are blocked now, so we don't get
3065 * a surprise exception before we get around to executing the
3068 if (tso->blocked_exceptions == NULL) {
3069 tso->blocked_exceptions = END_TSO_QUEUE;
3072 /* Put the newly-built THUNK on top of the stack, ready to execute
3073 * when the thread restarts.
3076 sp[-1] = (W_)&stg_enter_info;
3078 tso->what_next = ThreadRunGHC;
3079 IF_DEBUG(sanity, checkTSO(tso));
3088 // First build an AP_STACK consisting of the stack chunk above the
3089 // current update frame, with the top word on the stack as the
3092 words = frame - sp - 1;
3093 ap = (StgAP_STACK *)allocate(PAP_sizeW(words));
3096 ap->fun = (StgClosure *)sp[0];
3098 for(i=0; i < (nat)words; ++i) {
3099 ap->payload[i] = (StgClosure *)*sp++;
3102 SET_HDR(ap,&stg_AP_STACK_info,
3103 ((StgClosure *)frame)->header.prof.ccs /* ToDo */);
3104 TICK_ALLOC_UP_THK(words+1,0);
3107 fprintf(stderr, "sched: Updating ");
3108 printPtr((P_)((StgUpdateFrame *)frame)->updatee);
3109 fprintf(stderr, " with ");
3110 printObj((StgClosure *)ap);
3113 // Replace the updatee with an indirection - happily
3114 // this will also wake up any threads currently
3115 // waiting on the result.
3117 // Warning: if we're in a loop, more than one update frame on
3118 // the stack may point to the same object. Be careful not to
3119 // overwrite an IND_OLDGEN in this case, because we'll screw
3120 // up the mutable lists. To be on the safe side, don't
3121 // overwrite any kind of indirection at all. See also
3122 // threadSqueezeStack in GC.c, where we have to make a similar
3125 if (!closure_IND(((StgUpdateFrame *)frame)->updatee)) {
3126 // revert the black hole
3127 UPD_IND_NOLOCK(((StgUpdateFrame *)frame)->updatee,ap);
3129 sp += sizeofW(StgUpdateFrame) - 1;
3130 sp[0] = (W_)ap; // push onto stack
3135 // We've stripped the entire stack, the thread is now dead.
3136 sp += sizeofW(StgStopFrame);
3137 tso->what_next = ThreadKilled;
3148 /* -----------------------------------------------------------------------------
3149 resurrectThreads is called after garbage collection on the list of
3150 threads found to be garbage. Each of these threads will be woken
3151 up and sent a signal: BlockedOnDeadMVar if the thread was blocked
3152 on an MVar, or NonTermination if the thread was blocked on a Black
3155 Locks: sched_mutex isn't held upon entry nor exit.
3156 -------------------------------------------------------------------------- */
3159 resurrectThreads( StgTSO *threads )
3163 for (tso = threads; tso != END_TSO_QUEUE; tso = next) {
3164 next = tso->global_link;
3165 tso->global_link = all_threads;
3167 IF_DEBUG(scheduler, sched_belch("resurrecting thread %d", tso->id));
3169 switch (tso->why_blocked) {
3171 case BlockedOnException:
3172 /* Called by GC - sched_mutex lock is currently held. */
3173 raiseAsync(tso,(StgClosure *)BlockedOnDeadMVar_closure);
3175 case BlockedOnBlackHole:
3176 raiseAsync(tso,(StgClosure *)NonTermination_closure);
3179 /* This might happen if the thread was blocked on a black hole
3180 * belonging to a thread that we've just woken up (raiseAsync
3181 * can wake up threads, remember...).
3185 barf("resurrectThreads: thread blocked in a strange way");
3190 /* -----------------------------------------------------------------------------
3191 * Blackhole detection: if we reach a deadlock, test whether any
3192 * threads are blocked on themselves. Any threads which are found to
3193 * be self-blocked get sent a NonTermination exception.
3195 * This is only done in a deadlock situation in order to avoid
3196 * performance overhead in the normal case.
3198 * Locks: sched_mutex is held upon entry and exit.
3199 * -------------------------------------------------------------------------- */
3202 detectBlackHoles( void )
3204 StgTSO *tso = all_threads;
3206 StgClosure *blocked_on;
3207 StgRetInfoTable *info;
3209 for (tso = all_threads; tso != END_TSO_QUEUE; tso = tso->global_link) {
3211 while (tso->what_next == ThreadRelocated) {
3213 ASSERT(get_itbl(tso)->type == TSO);
3216 if (tso->why_blocked != BlockedOnBlackHole) {
3219 blocked_on = tso->block_info.closure;
3221 frame = (StgClosure *)tso->sp;
3224 info = get_ret_itbl(frame);
3225 switch (info->i.type) {
3227 if (((StgUpdateFrame *)frame)->updatee == blocked_on) {
3228 /* We are blocking on one of our own computations, so
3229 * send this thread the NonTermination exception.
3232 sched_belch("thread %d is blocked on itself", tso->id));
3233 raiseAsync(tso, (StgClosure *)NonTermination_closure);
3237 frame = (StgClosure *) ((StgUpdateFrame *)frame + 1);
3243 // normal stack frames; do nothing except advance the pointer
3245 (StgPtr)frame += stack_frame_sizeW(frame);
3252 /* ----------------------------------------------------------------------------
3253 * Debugging: why is a thread blocked
3254 * [Also provides useful information when debugging threaded programs
3255 * at the Haskell source code level, so enable outside of DEBUG. --sof 7/02]
3256 ------------------------------------------------------------------------- */
3260 printThreadBlockage(StgTSO *tso)
3262 switch (tso->why_blocked) {
3264 fprintf(stderr,"is blocked on read from fd %d", tso->block_info.fd);
3266 case BlockedOnWrite:
3267 fprintf(stderr,"is blocked on write to fd %d", tso->block_info.fd);
3269 #if defined(mingw32_TARGET_OS)
3270 case BlockedOnDoProc:
3271 fprintf(stderr,"is blocked on proc (request: %d)", tso->block_info.async_result->reqID);
3274 case BlockedOnDelay:
3275 fprintf(stderr,"is blocked until %d", tso->block_info.target);
3278 fprintf(stderr,"is blocked on an MVar");
3280 case BlockedOnException:
3281 fprintf(stderr,"is blocked on delivering an exception to thread %d",
3282 tso->block_info.tso->id);
3284 case BlockedOnBlackHole:
3285 fprintf(stderr,"is blocked on a black hole");
3288 fprintf(stderr,"is not blocked");
3292 fprintf(stderr,"is blocked on global address; local FM_BQ is %p (%s)",
3293 tso->block_info.closure, info_type(tso->block_info.closure));
3295 case BlockedOnGA_NoSend:
3296 fprintf(stderr,"is blocked on global address (no send); local FM_BQ is %p (%s)",
3297 tso->block_info.closure, info_type(tso->block_info.closure));
3300 case BlockedOnCCall:
3301 fprintf(stderr,"is blocked on an external call");
3303 case BlockedOnCCall_NoUnblockExc:
3304 fprintf(stderr,"is blocked on an external call (exceptions were already blocked)");
3307 barf("printThreadBlockage: strange tso->why_blocked: %d for TSO %d (%d)",
3308 tso->why_blocked, tso->id, tso);
3314 printThreadStatus(StgTSO *tso)
3316 switch (tso->what_next) {
3318 fprintf(stderr,"has been killed");
3320 case ThreadComplete:
3321 fprintf(stderr,"has completed");
3324 printThreadBlockage(tso);
3329 printAllThreads(void)
3335 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3336 ullong_format_string(TIME_ON_PROC(CurrentProc),
3337 time_string, rtsFalse/*no commas!*/);
3339 fprintf(stderr, "all threads at [%s]:\n", time_string);
3341 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3342 ullong_format_string(CURRENT_TIME,
3343 time_string, rtsFalse/*no commas!*/);
3345 fprintf(stderr,"all threads at [%s]:\n", time_string);
3347 fprintf(stderr,"all threads:\n");
3350 for (t = all_threads; t != END_TSO_QUEUE; t = t->global_link) {
3351 fprintf(stderr, "\tthread %d @ %p ", t->id, (void *)t);
3352 label = lookupThreadLabel(t->id);
3353 if (label) fprintf(stderr,"[\"%s\"] ",(char *)label);
3354 printThreadStatus(t);
3355 fprintf(stderr,"\n");
3362 Print a whole blocking queue attached to node (debugging only).
3366 print_bq (StgClosure *node)
3368 StgBlockingQueueElement *bqe;
3372 fprintf(stderr,"## BQ of closure %p (%s): ",
3373 node, info_type(node));
3375 /* should cover all closures that may have a blocking queue */
3376 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3377 get_itbl(node)->type == FETCH_ME_BQ ||
3378 get_itbl(node)->type == RBH ||
3379 get_itbl(node)->type == MVAR);
3381 ASSERT(node!=(StgClosure*)NULL); // sanity check
3383 print_bqe(((StgBlockingQueue*)node)->blocking_queue);
3387 Print a whole blocking queue starting with the element bqe.
3390 print_bqe (StgBlockingQueueElement *bqe)
3395 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3397 for (end = (bqe==END_BQ_QUEUE);
3398 !end; // iterate until bqe points to a CONSTR
3399 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE),
3400 bqe = end ? END_BQ_QUEUE : bqe->link) {
3401 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3402 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3403 /* types of closures that may appear in a blocking queue */
3404 ASSERT(get_itbl(bqe)->type == TSO ||
3405 get_itbl(bqe)->type == BLOCKED_FETCH ||
3406 get_itbl(bqe)->type == CONSTR);
3407 /* only BQs of an RBH end with an RBH_Save closure */
3408 //ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3410 switch (get_itbl(bqe)->type) {
3412 fprintf(stderr," TSO %u (%x),",
3413 ((StgTSO *)bqe)->id, ((StgTSO *)bqe));
3416 fprintf(stderr," BF (node=%p, ga=((%x, %d, %x)),",
3417 ((StgBlockedFetch *)bqe)->node,
3418 ((StgBlockedFetch *)bqe)->ga.payload.gc.gtid,
3419 ((StgBlockedFetch *)bqe)->ga.payload.gc.slot,
3420 ((StgBlockedFetch *)bqe)->ga.weight);
3423 fprintf(stderr," %s (IP %p),",
3424 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3425 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3426 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3427 "RBH_Save_?"), get_itbl(bqe));
3430 barf("Unexpected closure type %s in blocking queue", // of %p (%s)",
3431 info_type((StgClosure *)bqe)); // , node, info_type(node));
3435 fputc('\n', stderr);
3437 # elif defined(GRAN)
3439 print_bq (StgClosure *node)
3441 StgBlockingQueueElement *bqe;
3442 PEs node_loc, tso_loc;
3445 /* should cover all closures that may have a blocking queue */
3446 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3447 get_itbl(node)->type == FETCH_ME_BQ ||
3448 get_itbl(node)->type == RBH);
3450 ASSERT(node!=(StgClosure*)NULL); // sanity check
3451 node_loc = where_is(node);
3453 fprintf(stderr,"## BQ of closure %p (%s) on [PE %d]: ",
3454 node, info_type(node), node_loc);
3457 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3459 for (bqe = ((StgBlockingQueue*)node)->blocking_queue, end = (bqe==END_BQ_QUEUE);
3460 !end; // iterate until bqe points to a CONSTR
3461 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE), bqe = end ? END_BQ_QUEUE : bqe->link) {
3462 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3463 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3464 /* types of closures that may appear in a blocking queue */
3465 ASSERT(get_itbl(bqe)->type == TSO ||
3466 get_itbl(bqe)->type == CONSTR);
3467 /* only BQs of an RBH end with an RBH_Save closure */
3468 ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3470 tso_loc = where_is((StgClosure *)bqe);
3471 switch (get_itbl(bqe)->type) {
3473 fprintf(stderr," TSO %d (%p) on [PE %d],",
3474 ((StgTSO *)bqe)->id, (StgTSO *)bqe, tso_loc);
3477 fprintf(stderr," %s (IP %p),",
3478 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3479 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3480 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3481 "RBH_Save_?"), get_itbl(bqe));
3484 barf("Unexpected closure type %s in blocking queue of %p (%s)",
3485 info_type((StgClosure *)bqe), node, info_type(node));
3489 fputc('\n', stderr);
3493 Nice and easy: only TSOs on the blocking queue
3496 print_bq (StgClosure *node)
3500 ASSERT(node!=(StgClosure*)NULL); // sanity check
3501 for (tso = ((StgBlockingQueue*)node)->blocking_queue;
3502 tso != END_TSO_QUEUE;
3504 ASSERT(tso!=NULL && tso!=END_TSO_QUEUE); // sanity check
3505 ASSERT(get_itbl(tso)->type == TSO); // guess what, sanity check
3506 fprintf(stderr," TSO %d (%p),", tso->id, tso);
3508 fputc('\n', stderr);
3519 for (i=0, tso=run_queue_hd;
3520 tso != END_TSO_QUEUE;
3529 sched_belch(char *s, ...)
3533 #ifdef RTS_SUPPORTS_THREADS
3534 fprintf(stderr, "sched (task %p): ", osThreadId());
3536 fprintf(stderr, "== ");
3538 fprintf(stderr, "sched: ");
3540 vfprintf(stderr, s, ap);
3541 fprintf(stderr, "\n");