1 /* ---------------------------------------------------------------------------
2 * $Id: Schedule.c,v 1.195 2004/04/13 13:43:11 simonmar 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 startTask(taskStart);
285 /* ---------------------------------------------------------------------------
286 Main scheduling loop.
288 We use round-robin scheduling, each thread returning to the
289 scheduler loop when one of these conditions is detected:
292 * timer expires (thread yields)
297 Locking notes: we acquire the scheduler lock once at the beginning
298 of the scheduler loop, and release it when
300 * running a thread, or
301 * waiting for work, or
302 * waiting for a GC to complete.
305 In a GranSim setup this loop iterates over the global event queue.
306 This revolves around the global event queue, which determines what
307 to do next. Therefore, it's more complicated than either the
308 concurrent or the parallel (GUM) setup.
311 GUM iterates over incoming messages.
312 It starts with nothing to do (thus CurrentTSO == END_TSO_QUEUE),
313 and sends out a fish whenever it has nothing to do; in-between
314 doing the actual reductions (shared code below) it processes the
315 incoming messages and deals with delayed operations
316 (see PendingFetches).
317 This is not the ugliest code you could imagine, but it's bloody close.
319 ------------------------------------------------------------------------ */
321 schedule( StgMainThread *mainThread USED_WHEN_RTS_SUPPORTS_THREADS,
322 Capability *initialCapability )
326 StgThreadReturnCode ret;
334 rtsBool receivedFinish = rtsFalse;
336 nat tp_size, sp_size; // stats only
339 rtsBool was_interrupted = rtsFalse;
340 StgTSOWhatNext prev_what_next;
342 // Pre-condition: sched_mutex is held.
343 // We might have a capability, passed in as initialCapability.
344 cap = initialCapability;
346 #if defined(RTS_SUPPORTS_THREADS)
348 // in the threaded case, the capability is either passed in via the
349 // initialCapability parameter, or initialized inside the scheduler
353 sched_belch("### NEW SCHEDULER LOOP (main thr: %p, cap: %p)",
354 mainThread, initialCapability);
357 // simply initialise it in the non-threaded case
358 grabCapability(&cap);
362 /* set up first event to get things going */
363 /* ToDo: assign costs for system setup and init MainTSO ! */
364 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
366 CurrentTSO, (StgClosure*)NULL, (rtsSpark*)NULL);
369 fprintf(stderr, "GRAN: Init CurrentTSO (in schedule) = %p\n", CurrentTSO);
370 G_TSO(CurrentTSO, 5));
372 if (RtsFlags.GranFlags.Light) {
373 /* Save current time; GranSim Light only */
374 CurrentTSO->gran.clock = CurrentTime[CurrentProc];
377 event = get_next_event();
379 while (event!=(rtsEvent*)NULL) {
380 /* Choose the processor with the next event */
381 CurrentProc = event->proc;
382 CurrentTSO = event->tso;
386 while (!receivedFinish) { /* set by processMessages */
387 /* when receiving PP_FINISH message */
389 #else // everything except GRAN and PAR
395 IF_DEBUG(scheduler, printAllThreads());
397 #if defined(RTS_SUPPORTS_THREADS)
398 // Yield the capability to higher-priority tasks if necessary.
401 yieldCapability(&cap);
404 // If we do not currently hold a capability, we wait for one
407 waitForCapability(&sched_mutex, &cap,
408 mainThread ? &mainThread->bound_thread_cond : NULL);
411 // We now have a capability...
415 // If we're interrupted (the user pressed ^C, or some other
416 // termination condition occurred), kill all the currently running
420 IF_DEBUG(scheduler, sched_belch("interrupted"));
421 interrupted = rtsFalse;
422 was_interrupted = rtsTrue;
423 #if defined(RTS_SUPPORTS_THREADS)
424 // In the threaded RTS, deadlock detection doesn't work,
425 // so just exit right away.
426 prog_belch("interrupted");
427 releaseCapability(cap);
428 RELEASE_LOCK(&sched_mutex);
429 shutdownHaskellAndExit(EXIT_SUCCESS);
435 #if defined(RTS_USER_SIGNALS)
436 // check for signals each time around the scheduler
437 if (signals_pending()) {
438 RELEASE_LOCK(&sched_mutex); /* ToDo: kill */
439 startSignalHandlers();
440 ACQUIRE_LOCK(&sched_mutex);
445 // Check whether any waiting threads need to be woken up. If the
446 // run queue is empty, and there are no other tasks running, we
447 // can wait indefinitely for something to happen.
449 if ( !EMPTY_QUEUE(blocked_queue_hd) || !EMPTY_QUEUE(sleeping_queue)
450 #if defined(RTS_SUPPORTS_THREADS)
455 awaitEvent( EMPTY_RUN_QUEUE() );
457 // we can be interrupted while waiting for I/O...
458 if (interrupted) continue;
461 * Detect deadlock: when we have no threads to run, there are no
462 * threads waiting on I/O or sleeping, and all the other tasks are
463 * waiting for work, we must have a deadlock of some description.
465 * We first try to find threads blocked on themselves (ie. black
466 * holes), and generate NonTermination exceptions where necessary.
468 * If no threads are black holed, we have a deadlock situation, so
469 * inform all the main threads.
471 #if !defined(PAR) && !defined(RTS_SUPPORTS_THREADS)
472 if ( EMPTY_THREAD_QUEUES() )
474 IF_DEBUG(scheduler, sched_belch("deadlocked, forcing major GC..."));
475 // Garbage collection can release some new threads due to
476 // either (a) finalizers or (b) threads resurrected because
477 // they are about to be send BlockedOnDeadMVar. Any threads
478 // thus released will be immediately runnable.
479 GarbageCollect(GetRoots,rtsTrue);
481 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
484 sched_belch("still deadlocked, checking for black holes..."));
487 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
489 #if defined(RTS_USER_SIGNALS)
490 /* If we have user-installed signal handlers, then wait
491 * for signals to arrive rather then bombing out with a
494 if ( anyUserHandlers() ) {
496 sched_belch("still deadlocked, waiting for signals..."));
500 // we might be interrupted...
501 if (interrupted) { continue; }
503 if (signals_pending()) {
504 RELEASE_LOCK(&sched_mutex);
505 startSignalHandlers();
506 ACQUIRE_LOCK(&sched_mutex);
508 ASSERT(!EMPTY_RUN_QUEUE());
513 /* Probably a real deadlock. Send the current main thread the
514 * Deadlock exception (or in the SMP build, send *all* main
515 * threads the deadlock exception, since none of them can make
521 switch (m->tso->why_blocked) {
522 case BlockedOnBlackHole:
523 case BlockedOnException:
525 raiseAsync(m->tso, (StgClosure *)NonTermination_closure);
528 barf("deadlock: main thread blocked in a strange way");
534 #elif defined(RTS_SUPPORTS_THREADS)
535 // ToDo: add deadlock detection in threaded RTS
537 // ToDo: add deadlock detection in GUM (similar to SMP) -- HWL
540 #if defined(RTS_SUPPORTS_THREADS)
541 if ( EMPTY_RUN_QUEUE() ) {
542 continue; // nothing to do
547 if (RtsFlags.GranFlags.Light)
548 GranSimLight_enter_system(event, &ActiveTSO); // adjust ActiveTSO etc
550 /* adjust time based on time-stamp */
551 if (event->time > CurrentTime[CurrentProc] &&
552 event->evttype != ContinueThread)
553 CurrentTime[CurrentProc] = event->time;
555 /* Deal with the idle PEs (may issue FindWork or MoveSpark events) */
556 if (!RtsFlags.GranFlags.Light)
559 IF_DEBUG(gran, fprintf(stderr, "GRAN: switch by event-type\n"));
561 /* main event dispatcher in GranSim */
562 switch (event->evttype) {
563 /* Should just be continuing execution */
565 IF_DEBUG(gran, fprintf(stderr, "GRAN: doing ContinueThread\n"));
566 /* ToDo: check assertion
567 ASSERT(run_queue_hd != (StgTSO*)NULL &&
568 run_queue_hd != END_TSO_QUEUE);
570 /* Ignore ContinueThreads for fetching threads (if synchr comm) */
571 if (!RtsFlags.GranFlags.DoAsyncFetch &&
572 procStatus[CurrentProc]==Fetching) {
573 belch("ghuH: Spurious ContinueThread while Fetching ignored; TSO %d (%p) [PE %d]",
574 CurrentTSO->id, CurrentTSO, CurrentProc);
577 /* Ignore ContinueThreads for completed threads */
578 if (CurrentTSO->what_next == ThreadComplete) {
579 belch("ghuH: found a ContinueThread event for completed thread %d (%p) [PE %d] (ignoring ContinueThread)",
580 CurrentTSO->id, CurrentTSO, CurrentProc);
583 /* Ignore ContinueThreads for threads that are being migrated */
584 if (PROCS(CurrentTSO)==Nowhere) {
585 belch("ghuH: trying to run the migrating TSO %d (%p) [PE %d] (ignoring ContinueThread)",
586 CurrentTSO->id, CurrentTSO, CurrentProc);
589 /* The thread should be at the beginning of the run queue */
590 if (CurrentTSO!=run_queue_hds[CurrentProc]) {
591 belch("ghuH: TSO %d (%p) [PE %d] is not at the start of the run_queue when doing a ContinueThread",
592 CurrentTSO->id, CurrentTSO, CurrentProc);
593 break; // run the thread anyway
596 new_event(proc, proc, CurrentTime[proc],
598 (StgTSO*)NULL, (StgClosure*)NULL, (rtsSpark*)NULL);
600 */ /* Catches superfluous CONTINUEs -- should be unnecessary */
601 break; // now actually run the thread; DaH Qu'vam yImuHbej
604 do_the_fetchnode(event);
605 goto next_thread; /* handle next event in event queue */
608 do_the_globalblock(event);
609 goto next_thread; /* handle next event in event queue */
612 do_the_fetchreply(event);
613 goto next_thread; /* handle next event in event queue */
615 case UnblockThread: /* Move from the blocked queue to the tail of */
616 do_the_unblock(event);
617 goto next_thread; /* handle next event in event queue */
619 case ResumeThread: /* Move from the blocked queue to the tail of */
620 /* the runnable queue ( i.e. Qu' SImqa'lu') */
621 event->tso->gran.blocktime +=
622 CurrentTime[CurrentProc] - event->tso->gran.blockedat;
623 do_the_startthread(event);
624 goto next_thread; /* handle next event in event queue */
627 do_the_startthread(event);
628 goto next_thread; /* handle next event in event queue */
631 do_the_movethread(event);
632 goto next_thread; /* handle next event in event queue */
635 do_the_movespark(event);
636 goto next_thread; /* handle next event in event queue */
639 do_the_findwork(event);
640 goto next_thread; /* handle next event in event queue */
643 barf("Illegal event type %u\n", event->evttype);
646 /* This point was scheduler_loop in the old RTS */
648 IF_DEBUG(gran, belch("GRAN: after main switch"));
650 TimeOfLastEvent = CurrentTime[CurrentProc];
651 TimeOfNextEvent = get_time_of_next_event();
652 IgnoreEvents=(TimeOfNextEvent==0); // HWL HACK
653 // CurrentTSO = ThreadQueueHd;
655 IF_DEBUG(gran, belch("GRAN: time of next event is: %ld",
658 if (RtsFlags.GranFlags.Light)
659 GranSimLight_leave_system(event, &ActiveTSO);
661 EndOfTimeSlice = CurrentTime[CurrentProc]+RtsFlags.GranFlags.time_slice;
664 belch("GRAN: end of time-slice is %#lx", EndOfTimeSlice));
666 /* in a GranSim setup the TSO stays on the run queue */
668 /* Take a thread from the run queue. */
669 POP_RUN_QUEUE(t); // take_off_run_queue(t);
672 fprintf(stderr, "GRAN: About to run current thread, which is\n");
675 context_switch = 0; // turned on via GranYield, checking events and time slice
678 DumpGranEvent(GR_SCHEDULE, t));
680 procStatus[CurrentProc] = Busy;
683 if (PendingFetches != END_BF_QUEUE) {
687 /* ToDo: phps merge with spark activation above */
688 /* check whether we have local work and send requests if we have none */
689 if (EMPTY_RUN_QUEUE()) { /* no runnable threads */
690 /* :-[ no local threads => look out for local sparks */
691 /* the spark pool for the current PE */
692 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
693 if (advisory_thread_count < RtsFlags.ParFlags.maxThreads &&
694 pool->hd < pool->tl) {
696 * ToDo: add GC code check that we really have enough heap afterwards!!
698 * If we're here (no runnable threads) and we have pending
699 * sparks, we must have a space problem. Get enough space
700 * to turn one of those pending sparks into a
704 spark = findSpark(rtsFalse); /* get a spark */
705 if (spark != (rtsSpark) NULL) {
706 tso = activateSpark(spark); /* turn the spark into a thread */
707 IF_PAR_DEBUG(schedule,
708 belch("==== schedule: Created TSO %d (%p); %d threads active",
709 tso->id, tso, advisory_thread_count));
711 if (tso==END_TSO_QUEUE) { /* failed to activate spark->back to loop */
712 belch("==^^ failed to activate spark");
714 } /* otherwise fall through & pick-up new tso */
716 IF_PAR_DEBUG(verbose,
717 belch("==^^ no local sparks (spark pool contains only NFs: %d)",
718 spark_queue_len(pool)));
723 /* If we still have no work we need to send a FISH to get a spark
726 if (EMPTY_RUN_QUEUE()) {
727 /* =8-[ no local sparks => look for work on other PEs */
729 * We really have absolutely no work. Send out a fish
730 * (there may be some out there already), and wait for
731 * something to arrive. We clearly can't run any threads
732 * until a SCHEDULE or RESUME arrives, and so that's what
733 * we're hoping to see. (Of course, we still have to
734 * respond to other types of messages.)
736 TIME now = msTime() /*CURRENT_TIME*/;
737 IF_PAR_DEBUG(verbose,
738 belch("-- now=%ld", now));
739 IF_PAR_DEBUG(verbose,
740 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
741 (last_fish_arrived_at!=0 &&
742 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay > now)) {
743 belch("--$$ delaying FISH until %ld (last fish %ld, delay %ld, now %ld)",
744 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay,
745 last_fish_arrived_at,
746 RtsFlags.ParFlags.fishDelay, now);
749 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
750 (last_fish_arrived_at==0 ||
751 (last_fish_arrived_at+RtsFlags.ParFlags.fishDelay <= now))) {
752 /* outstandingFishes is set in sendFish, processFish;
753 avoid flooding system with fishes via delay */
755 sendFish(pe, mytid, NEW_FISH_AGE, NEW_FISH_HISTORY,
758 // Global statistics: count no. of fishes
759 if (RtsFlags.ParFlags.ParStats.Global &&
760 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
761 globalParStats.tot_fish_mess++;
765 receivedFinish = processMessages();
768 } else if (PacketsWaiting()) { /* Look for incoming messages */
769 receivedFinish = processMessages();
772 /* Now we are sure that we have some work available */
773 ASSERT(run_queue_hd != END_TSO_QUEUE);
775 /* Take a thread from the run queue, if we have work */
776 POP_RUN_QUEUE(t); // take_off_run_queue(END_TSO_QUEUE);
777 IF_DEBUG(sanity,checkTSO(t));
779 /* ToDo: write something to the log-file
780 if (RTSflags.ParFlags.granSimStats && !sameThread)
781 DumpGranEvent(GR_SCHEDULE, RunnableThreadsHd);
785 /* the spark pool for the current PE */
786 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
789 belch("--=^ %d threads, %d sparks on [%#x]",
790 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
793 if (0 && RtsFlags.ParFlags.ParStats.Full &&
794 t && LastTSO && t->id != LastTSO->id &&
795 LastTSO->why_blocked == NotBlocked &&
796 LastTSO->what_next != ThreadComplete) {
797 // if previously scheduled TSO not blocked we have to record the context switch
798 DumpVeryRawGranEvent(TimeOfLastYield, CURRENT_PROC, CURRENT_PROC,
799 GR_DESCHEDULE, LastTSO, (StgClosure *)NULL, 0, 0);
802 if (RtsFlags.ParFlags.ParStats.Full &&
803 (emitSchedule /* forced emit */ ||
804 (t && LastTSO && t->id != LastTSO->id))) {
806 we are running a different TSO, so write a schedule event to log file
807 NB: If we use fair scheduling we also have to write a deschedule
808 event for LastTSO; with unfair scheduling we know that the
809 previous tso has blocked whenever we switch to another tso, so
810 we don't need it in GUM for now
812 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
813 GR_SCHEDULE, t, (StgClosure *)NULL, 0, 0);
814 emitSchedule = rtsFalse;
818 #else /* !GRAN && !PAR */
820 // grab a thread from the run queue
821 ASSERT(run_queue_hd != END_TSO_QUEUE);
824 // Sanity check the thread we're about to run. This can be
825 // expensive if there is lots of thread switching going on...
826 IF_DEBUG(sanity,checkTSO(t));
831 StgMainThread *m = t->main;
838 sched_belch("### Running thread %d in bound thread", t->id));
839 // yes, the Haskell thread is bound to the current native thread
844 sched_belch("### thread %d bound to another OS thread", t->id));
845 // no, bound to a different Haskell thread: pass to that thread
846 PUSH_ON_RUN_QUEUE(t);
847 passCapability(&m->bound_thread_cond);
853 if(mainThread != NULL)
854 // The thread we want to run is bound.
857 sched_belch("### this OS thread cannot run thread %d", t->id));
858 // no, the current native thread is bound to a different
859 // Haskell thread, so pass it to any worker thread
860 PUSH_ON_RUN_QUEUE(t);
861 passCapabilityToWorker();
868 cap->r.rCurrentTSO = t;
870 /* context switches are now initiated by the timer signal, unless
871 * the user specified "context switch as often as possible", with
874 if ((RtsFlags.ConcFlags.ctxtSwitchTicks == 0
875 && (run_queue_hd != END_TSO_QUEUE
876 || blocked_queue_hd != END_TSO_QUEUE
877 || sleeping_queue != END_TSO_QUEUE)))
884 RELEASE_LOCK(&sched_mutex);
886 IF_DEBUG(scheduler, sched_belch("-->> running thread %ld %s ...",
887 t->id, whatNext_strs[t->what_next]));
890 startHeapProfTimer();
893 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
894 /* Run the current thread
896 prev_what_next = t->what_next;
897 switch (prev_what_next) {
900 /* Thread already finished, return to scheduler. */
901 ret = ThreadFinished;
904 errno = t->saved_errno;
905 ret = StgRun((StgFunPtr) stg_returnToStackTop, &cap->r);
906 t->saved_errno = errno;
908 case ThreadInterpret:
909 ret = interpretBCO(cap);
912 barf("schedule: invalid what_next field");
914 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
916 /* Costs for the scheduler are assigned to CCS_SYSTEM */
922 ACQUIRE_LOCK(&sched_mutex);
924 #ifdef RTS_SUPPORTS_THREADS
925 IF_DEBUG(scheduler,fprintf(stderr,"sched (task %p): ", osThreadId()););
926 #elif !defined(GRAN) && !defined(PAR)
927 IF_DEBUG(scheduler,fprintf(stderr,"sched: "););
929 t = cap->r.rCurrentTSO;
932 /* HACK 675: if the last thread didn't yield, make sure to print a
933 SCHEDULE event to the log file when StgRunning the next thread, even
934 if it is the same one as before */
936 TimeOfLastYield = CURRENT_TIME;
942 IF_DEBUG(gran, DumpGranEvent(GR_DESCHEDULE, t));
943 globalGranStats.tot_heapover++;
945 globalParStats.tot_heapover++;
948 // did the task ask for a large block?
949 if (cap->r.rHpAlloc > BLOCK_SIZE_W) {
950 // if so, get one and push it on the front of the nursery.
954 blocks = (nat)BLOCK_ROUND_UP(cap->r.rHpAlloc * sizeof(W_)) / BLOCK_SIZE;
956 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped: requesting a large block (size %d)",
957 t->id, whatNext_strs[t->what_next], blocks));
959 // don't do this if it would push us over the
960 // alloc_blocks_lim limit; we'll GC first.
961 if (alloc_blocks + blocks < alloc_blocks_lim) {
963 alloc_blocks += blocks;
964 bd = allocGroup( blocks );
966 // link the new group into the list
967 bd->link = cap->r.rCurrentNursery;
968 bd->u.back = cap->r.rCurrentNursery->u.back;
969 if (cap->r.rCurrentNursery->u.back != NULL) {
970 cap->r.rCurrentNursery->u.back->link = bd;
972 ASSERT(g0s0->blocks == cap->r.rCurrentNursery &&
973 g0s0->blocks == cap->r.rNursery);
974 cap->r.rNursery = g0s0->blocks = bd;
976 cap->r.rCurrentNursery->u.back = bd;
978 // initialise it as a nursery block. We initialise the
979 // step, gen_no, and flags field of *every* sub-block in
980 // this large block, because this is easier than making
981 // sure that we always find the block head of a large
982 // block whenever we call Bdescr() (eg. evacuate() and
983 // isAlive() in the GC would both have to do this, at
987 for (x = bd; x < bd + blocks; x++) {
994 // don't forget to update the block count in g0s0.
995 g0s0->n_blocks += blocks;
996 // This assert can be a killer if the app is doing lots
997 // of large block allocations.
998 ASSERT(countBlocks(g0s0->blocks) == g0s0->n_blocks);
1000 // now update the nursery to point to the new block
1001 cap->r.rCurrentNursery = bd;
1003 // we might be unlucky and have another thread get on the
1004 // run queue before us and steal the large block, but in that
1005 // case the thread will just end up requesting another large
1007 PUSH_ON_RUN_QUEUE(t);
1012 /* make all the running tasks block on a condition variable,
1013 * maybe set context_switch and wait till they all pile in,
1014 * then have them wait on a GC condition variable.
1016 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped: HeapOverflow",
1017 t->id, whatNext_strs[t->what_next]));
1020 ASSERT(!is_on_queue(t,CurrentProc));
1022 /* Currently we emit a DESCHEDULE event before GC in GUM.
1023 ToDo: either add separate event to distinguish SYSTEM time from rest
1024 or just nuke this DESCHEDULE (and the following SCHEDULE) */
1025 if (0 && RtsFlags.ParFlags.ParStats.Full) {
1026 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1027 GR_DESCHEDULE, t, (StgClosure *)NULL, 0, 0);
1028 emitSchedule = rtsTrue;
1032 ready_to_gc = rtsTrue;
1033 context_switch = 1; /* stop other threads ASAP */
1034 PUSH_ON_RUN_QUEUE(t);
1035 /* actual GC is done at the end of the while loop */
1041 DumpGranEvent(GR_DESCHEDULE, t));
1042 globalGranStats.tot_stackover++;
1045 // DumpGranEvent(GR_DESCHEDULE, t);
1046 globalParStats.tot_stackover++;
1048 IF_DEBUG(scheduler,belch("--<< thread %ld (%s) stopped, StackOverflow",
1049 t->id, whatNext_strs[t->what_next]));
1050 /* just adjust the stack for this thread, then pop it back
1055 /* enlarge the stack */
1056 StgTSO *new_t = threadStackOverflow(t);
1058 /* This TSO has moved, so update any pointers to it from the
1059 * main thread stack. It better not be on any other queues...
1060 * (it shouldn't be).
1062 if (t->main != NULL) {
1063 t->main->tso = new_t;
1065 PUSH_ON_RUN_QUEUE(new_t);
1069 case ThreadYielding:
1072 DumpGranEvent(GR_DESCHEDULE, t));
1073 globalGranStats.tot_yields++;
1076 // DumpGranEvent(GR_DESCHEDULE, t);
1077 globalParStats.tot_yields++;
1079 /* put the thread back on the run queue. Then, if we're ready to
1080 * GC, check whether this is the last task to stop. If so, wake
1081 * up the GC thread. getThread will block during a GC until the
1085 if (t->what_next != prev_what_next) {
1086 belch("--<< thread %ld (%s) stopped to switch evaluators",
1087 t->id, whatNext_strs[t->what_next]);
1089 belch("--<< thread %ld (%s) stopped, yielding",
1090 t->id, whatNext_strs[t->what_next]);
1095 //belch("&& Doing sanity check on yielding TSO %ld.", t->id);
1097 ASSERT(t->link == END_TSO_QUEUE);
1099 // Shortcut if we're just switching evaluators: don't bother
1100 // doing stack squeezing (which can be expensive), just run the
1102 if (t->what_next != prev_what_next) {
1109 ASSERT(!is_on_queue(t,CurrentProc));
1112 //belch("&& Doing sanity check on all ThreadQueues (and their TSOs).");
1113 checkThreadQsSanity(rtsTrue));
1117 if (RtsFlags.ParFlags.doFairScheduling) {
1118 /* this does round-robin scheduling; good for concurrency */
1119 APPEND_TO_RUN_QUEUE(t);
1121 /* this does unfair scheduling; good for parallelism */
1122 PUSH_ON_RUN_QUEUE(t);
1125 // this does round-robin scheduling; good for concurrency
1126 APPEND_TO_RUN_QUEUE(t);
1130 /* add a ContinueThread event to actually process the thread */
1131 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1133 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1135 belch("GRAN: eventq and runnableq after adding yielded thread to queue again:");
1144 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: ",
1145 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)));
1146 if (t->block_info.closure!=(StgClosure*)NULL) print_bq(t->block_info.closure));
1148 // ??? needed; should emit block before
1150 DumpGranEvent(GR_DESCHEDULE, t));
1151 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1154 ASSERT(procStatus[CurrentProc]==Busy ||
1155 ((procStatus[CurrentProc]==Fetching) &&
1156 (t->block_info.closure!=(StgClosure*)NULL)));
1157 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1158 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1159 procStatus[CurrentProc]==Fetching))
1160 procStatus[CurrentProc] = Idle;
1164 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p with BQ: ",
1165 t->id, t, whatNext_strs[t->what_next], t->block_info.closure));
1168 if (t->block_info.closure!=(StgClosure*)NULL)
1169 print_bq(t->block_info.closure));
1171 /* Send a fetch (if BlockedOnGA) and dump event to log file */
1174 /* whatever we schedule next, we must log that schedule */
1175 emitSchedule = rtsTrue;
1178 /* don't need to do anything. Either the thread is blocked on
1179 * I/O, in which case we'll have called addToBlockedQueue
1180 * previously, or it's blocked on an MVar or Blackhole, in which
1181 * case it'll be on the relevant queue already.
1184 fprintf(stderr, "--<< thread %d (%s) stopped: ",
1185 t->id, whatNext_strs[t->what_next]);
1186 printThreadBlockage(t);
1187 fprintf(stderr, "\n"));
1190 /* Only for dumping event to log file
1191 ToDo: do I need this in GranSim, too?
1198 case ThreadFinished:
1199 /* Need to check whether this was a main thread, and if so, signal
1200 * the task that started it with the return value. If we have no
1201 * more main threads, we probably need to stop all the tasks until
1204 /* We also end up here if the thread kills itself with an
1205 * uncaught exception, see Exception.hc.
1207 IF_DEBUG(scheduler,belch("--++ thread %d (%s) finished",
1208 t->id, whatNext_strs[t->what_next]));
1210 endThread(t, CurrentProc); // clean-up the thread
1212 /* For now all are advisory -- HWL */
1213 //if(t->priority==AdvisoryPriority) ??
1214 advisory_thread_count--;
1217 if(t->dist.priority==RevalPriority)
1221 if (RtsFlags.ParFlags.ParStats.Full &&
1222 !RtsFlags.ParFlags.ParStats.Suppressed)
1223 DumpEndEvent(CURRENT_PROC, t, rtsFalse /* not mandatory */);
1227 // Check whether the thread that just completed was a main
1228 // thread, and if so return with the result.
1230 // There is an assumption here that all thread completion goes
1231 // through this point; we need to make sure that if a thread
1232 // ends up in the ThreadKilled state, that it stays on the run
1233 // queue so it can be dealt with here.
1236 #if defined(RTS_SUPPORTS_THREADS)
1239 mainThread->tso == t
1243 // We are a bound thread: this must be our thread that just
1245 ASSERT(mainThread->tso == t);
1247 if (t->what_next == ThreadComplete) {
1248 if (mainThread->ret) {
1249 // NOTE: return val is tso->sp[1] (see StgStartup.hc)
1250 *(mainThread->ret) = (StgClosure *)mainThread->tso->sp[1];
1252 mainThread->stat = Success;
1254 if (mainThread->ret) {
1255 *(mainThread->ret) = NULL;
1257 if (was_interrupted) {
1258 mainThread->stat = Interrupted;
1260 mainThread->stat = Killed;
1264 removeThreadLabel((StgWord)mainThread->tso->id);
1266 if (mainThread->prev == NULL) {
1267 main_threads = mainThread->link;
1269 mainThread->prev->link = mainThread->link;
1271 if (mainThread->link != NULL) {
1272 mainThread->link->prev = NULL;
1274 releaseCapability(cap);
1278 #ifdef RTS_SUPPORTS_THREADS
1279 ASSERT(t->main == NULL);
1281 if (t->main != NULL) {
1282 // Must be a main thread that is not the topmost one. Leave
1283 // it on the run queue until the stack has unwound to the
1284 // point where we can deal with this. Leaving it on the run
1285 // queue also ensures that the garbage collector knows about
1286 // this thread and its return value (it gets dropped from the
1287 // all_threads list so there's no other way to find it).
1288 APPEND_TO_RUN_QUEUE(t);
1294 barf("schedule: invalid thread return code %d", (int)ret);
1298 // When we have +RTS -i0 and we're heap profiling, do a census at
1299 // every GC. This lets us get repeatable runs for debugging.
1300 if (performHeapProfile ||
1301 (RtsFlags.ProfFlags.profileInterval==0 &&
1302 RtsFlags.ProfFlags.doHeapProfile && ready_to_gc)) {
1303 GarbageCollect(GetRoots, rtsTrue);
1305 performHeapProfile = rtsFalse;
1306 ready_to_gc = rtsFalse; // we already GC'd
1311 /* everybody back, start the GC.
1312 * Could do it in this thread, or signal a condition var
1313 * to do it in another thread. Either way, we need to
1314 * broadcast on gc_pending_cond afterward.
1316 #if defined(RTS_SUPPORTS_THREADS)
1317 IF_DEBUG(scheduler,sched_belch("doing GC"));
1319 GarbageCollect(GetRoots,rtsFalse);
1320 ready_to_gc = rtsFalse;
1322 /* add a ContinueThread event to continue execution of current thread */
1323 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1325 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1327 fprintf(stderr, "GRAN: eventq and runnableq after Garbage collection:\n");
1335 IF_GRAN_DEBUG(unused,
1336 print_eventq(EventHd));
1338 event = get_next_event();
1341 /* ToDo: wait for next message to arrive rather than busy wait */
1344 } /* end of while(1) */
1346 IF_PAR_DEBUG(verbose,
1347 belch("== Leaving schedule() after having received Finish"));
1350 /* ---------------------------------------------------------------------------
1351 * rtsSupportsBoundThreads(): is the RTS built to support bound threads?
1352 * used by Control.Concurrent for error checking.
1353 * ------------------------------------------------------------------------- */
1356 rtsSupportsBoundThreads(void)
1365 /* ---------------------------------------------------------------------------
1366 * isThreadBound(tso): check whether tso is bound to an OS thread.
1367 * ------------------------------------------------------------------------- */
1370 isThreadBound(StgTSO* tso USED_IN_THREADED_RTS)
1373 return (tso->main != NULL);
1378 /* ---------------------------------------------------------------------------
1379 * Singleton fork(). Do not copy any running threads.
1380 * ------------------------------------------------------------------------- */
1382 #ifndef mingw32_TARGET_OS
1383 #define FORKPROCESS_PRIMOP_SUPPORTED
1386 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1388 deleteThreadImmediately(StgTSO *tso);
1391 forkProcess(HsStablePtr *entry
1392 #ifndef FORKPROCESS_PRIMOP_SUPPORTED
1397 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
1403 IF_DEBUG(scheduler,sched_belch("forking!"));
1404 rts_lock(); // This not only acquires sched_mutex, it also
1405 // makes sure that no other threads are running
1409 if (pid) { /* parent */
1411 /* just return the pid */
1415 } else { /* child */
1418 // delete all threads
1419 run_queue_hd = run_queue_tl = END_TSO_QUEUE;
1421 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1424 // don't allow threads to catch the ThreadKilled exception
1425 deleteThreadImmediately(t);
1428 // wipe the main thread list
1429 while((m = main_threads) != NULL) {
1430 main_threads = m->link;
1431 # ifdef THREADED_RTS
1432 closeCondition(&m->bound_thread_cond);
1437 # ifdef RTS_SUPPORTS_THREADS
1438 resetTaskManagerAfterFork(); // tell startTask() and friends that
1439 startingWorkerThread = rtsFalse; // we have no worker threads any more
1440 resetWorkerWakeupPipeAfterFork();
1443 rc = rts_evalStableIO(entry, NULL); // run the action
1444 rts_checkSchedStatus("forkProcess",rc);
1448 hs_exit(); // clean up and exit
1451 #else /* !FORKPROCESS_PRIMOP_SUPPORTED */
1452 barf("forkProcess#: primop not supported, sorry!\n");
1457 /* ---------------------------------------------------------------------------
1458 * deleteAllThreads(): kill all the live threads.
1460 * This is used when we catch a user interrupt (^C), before performing
1461 * any necessary cleanups and running finalizers.
1463 * Locks: sched_mutex held.
1464 * ------------------------------------------------------------------------- */
1467 deleteAllThreads ( void )
1470 IF_DEBUG(scheduler,sched_belch("deleting all threads"));
1471 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1472 next = t->global_link;
1476 // The run queue now contains a bunch of ThreadKilled threads. We
1477 // must not throw these away: the main thread(s) will be in there
1478 // somewhere, and the main scheduler loop has to deal with it.
1479 // Also, the run queue is the only thing keeping these threads from
1480 // being GC'd, and we don't want the "main thread has been GC'd" panic.
1482 ASSERT(blocked_queue_hd == END_TSO_QUEUE);
1483 ASSERT(sleeping_queue == END_TSO_QUEUE);
1486 /* startThread and insertThread are now in GranSim.c -- HWL */
1489 /* ---------------------------------------------------------------------------
1490 * Suspending & resuming Haskell threads.
1492 * When making a "safe" call to C (aka _ccall_GC), the task gives back
1493 * its capability before calling the C function. This allows another
1494 * task to pick up the capability and carry on running Haskell
1495 * threads. It also means that if the C call blocks, it won't lock
1498 * The Haskell thread making the C call is put to sleep for the
1499 * duration of the call, on the susepended_ccalling_threads queue. We
1500 * give out a token to the task, which it can use to resume the thread
1501 * on return from the C function.
1502 * ------------------------------------------------------------------------- */
1505 suspendThread( StgRegTable *reg,
1514 int saved_errno = errno;
1516 /* assume that *reg is a pointer to the StgRegTable part
1519 cap = (Capability *)((void *)((unsigned char*)reg - sizeof(StgFunTable)));
1521 ACQUIRE_LOCK(&sched_mutex);
1524 sched_belch("thread %d did a _ccall_gc (is_concurrent: %d)", cap->r.rCurrentTSO->id,concCall));
1526 // XXX this might not be necessary --SDM
1527 cap->r.rCurrentTSO->what_next = ThreadRunGHC;
1529 threadPaused(cap->r.rCurrentTSO);
1530 cap->r.rCurrentTSO->link = suspended_ccalling_threads;
1531 suspended_ccalling_threads = cap->r.rCurrentTSO;
1533 if(cap->r.rCurrentTSO->blocked_exceptions == NULL) {
1534 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall;
1535 cap->r.rCurrentTSO->blocked_exceptions = END_TSO_QUEUE;
1537 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall_NoUnblockExc;
1540 /* Use the thread ID as the token; it should be unique */
1541 tok = cap->r.rCurrentTSO->id;
1543 /* Hand back capability */
1544 releaseCapability(cap);
1546 #if defined(RTS_SUPPORTS_THREADS)
1547 /* Preparing to leave the RTS, so ensure there's a native thread/task
1548 waiting to take over.
1550 IF_DEBUG(scheduler, sched_belch("worker (token %d): leaving RTS", tok));
1553 /* Other threads _might_ be available for execution; signal this */
1555 RELEASE_LOCK(&sched_mutex);
1557 errno = saved_errno;
1562 resumeThread( StgInt tok,
1563 rtsBool concCall STG_UNUSED )
1565 StgTSO *tso, **prev;
1567 int saved_errno = errno;
1569 #if defined(RTS_SUPPORTS_THREADS)
1570 /* Wait for permission to re-enter the RTS with the result. */
1571 ACQUIRE_LOCK(&sched_mutex);
1572 waitForReturnCapability(&sched_mutex, &cap);
1574 IF_DEBUG(scheduler, sched_belch("worker (token %d): re-entering RTS", tok));
1576 grabCapability(&cap);
1579 /* Remove the thread off of the suspended list */
1580 prev = &suspended_ccalling_threads;
1581 for (tso = suspended_ccalling_threads;
1582 tso != END_TSO_QUEUE;
1583 prev = &tso->link, tso = tso->link) {
1584 if (tso->id == (StgThreadID)tok) {
1589 if (tso == END_TSO_QUEUE) {
1590 barf("resumeThread: thread not found");
1592 tso->link = END_TSO_QUEUE;
1594 if(tso->why_blocked == BlockedOnCCall) {
1595 awakenBlockedQueueNoLock(tso->blocked_exceptions);
1596 tso->blocked_exceptions = NULL;
1599 /* Reset blocking status */
1600 tso->why_blocked = NotBlocked;
1602 cap->r.rCurrentTSO = tso;
1603 RELEASE_LOCK(&sched_mutex);
1604 errno = saved_errno;
1609 /* ---------------------------------------------------------------------------
1611 * ------------------------------------------------------------------------ */
1612 static void unblockThread(StgTSO *tso);
1614 /* ---------------------------------------------------------------------------
1615 * Comparing Thread ids.
1617 * This is used from STG land in the implementation of the
1618 * instances of Eq/Ord for ThreadIds.
1619 * ------------------------------------------------------------------------ */
1622 cmp_thread(StgPtr tso1, StgPtr tso2)
1624 StgThreadID id1 = ((StgTSO *)tso1)->id;
1625 StgThreadID id2 = ((StgTSO *)tso2)->id;
1627 if (id1 < id2) return (-1);
1628 if (id1 > id2) return 1;
1632 /* ---------------------------------------------------------------------------
1633 * Fetching the ThreadID from an StgTSO.
1635 * This is used in the implementation of Show for ThreadIds.
1636 * ------------------------------------------------------------------------ */
1638 rts_getThreadId(StgPtr tso)
1640 return ((StgTSO *)tso)->id;
1645 labelThread(StgPtr tso, char *label)
1650 /* Caveat: Once set, you can only set the thread name to "" */
1651 len = strlen(label)+1;
1652 buf = stgMallocBytes(len * sizeof(char), "Schedule.c:labelThread()");
1653 strncpy(buf,label,len);
1654 /* Update will free the old memory for us */
1655 updateThreadLabel(((StgTSO *)tso)->id,buf);
1659 /* ---------------------------------------------------------------------------
1660 Create a new thread.
1662 The new thread starts with the given stack size. Before the
1663 scheduler can run, however, this thread needs to have a closure
1664 (and possibly some arguments) pushed on its stack. See
1665 pushClosure() in Schedule.h.
1667 createGenThread() and createIOThread() (in SchedAPI.h) are
1668 convenient packaged versions of this function.
1670 currently pri (priority) is only used in a GRAN setup -- HWL
1671 ------------------------------------------------------------------------ */
1673 /* currently pri (priority) is only used in a GRAN setup -- HWL */
1675 createThread(nat size, StgInt pri)
1678 createThread(nat size)
1685 /* First check whether we should create a thread at all */
1687 /* check that no more than RtsFlags.ParFlags.maxThreads threads are created */
1688 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads) {
1690 belch("{createThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1691 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1692 return END_TSO_QUEUE;
1698 ASSERT(!RtsFlags.GranFlags.Light || CurrentProc==0);
1701 // ToDo: check whether size = stack_size - TSO_STRUCT_SIZEW
1703 /* catch ridiculously small stack sizes */
1704 if (size < MIN_STACK_WORDS + TSO_STRUCT_SIZEW) {
1705 size = MIN_STACK_WORDS + TSO_STRUCT_SIZEW;
1708 stack_size = size - TSO_STRUCT_SIZEW;
1710 tso = (StgTSO *)allocate(size);
1711 TICK_ALLOC_TSO(stack_size, 0);
1713 SET_HDR(tso, &stg_TSO_info, CCS_SYSTEM);
1715 SET_GRAN_HDR(tso, ThisPE);
1718 // Always start with the compiled code evaluator
1719 tso->what_next = ThreadRunGHC;
1721 tso->id = next_thread_id++;
1722 tso->why_blocked = NotBlocked;
1723 tso->blocked_exceptions = NULL;
1725 tso->saved_errno = 0;
1728 tso->stack_size = stack_size;
1729 tso->max_stack_size = round_to_mblocks(RtsFlags.GcFlags.maxStkSize)
1731 tso->sp = (P_)&(tso->stack) + stack_size;
1734 tso->prof.CCCS = CCS_MAIN;
1737 /* put a stop frame on the stack */
1738 tso->sp -= sizeofW(StgStopFrame);
1739 SET_HDR((StgClosure*)tso->sp,(StgInfoTable *)&stg_stop_thread_info,CCS_SYSTEM);
1742 tso->link = END_TSO_QUEUE;
1743 /* uses more flexible routine in GranSim */
1744 insertThread(tso, CurrentProc);
1746 /* In a non-GranSim setup the pushing of a TSO onto the runq is separated
1752 if (RtsFlags.GranFlags.GranSimStats.Full)
1753 DumpGranEvent(GR_START,tso);
1755 if (RtsFlags.ParFlags.ParStats.Full)
1756 DumpGranEvent(GR_STARTQ,tso);
1757 /* HACk to avoid SCHEDULE
1761 /* Link the new thread on the global thread list.
1763 tso->global_link = all_threads;
1767 tso->dist.priority = MandatoryPriority; //by default that is...
1771 tso->gran.pri = pri;
1773 tso->gran.magic = TSO_MAGIC; // debugging only
1775 tso->gran.sparkname = 0;
1776 tso->gran.startedat = CURRENT_TIME;
1777 tso->gran.exported = 0;
1778 tso->gran.basicblocks = 0;
1779 tso->gran.allocs = 0;
1780 tso->gran.exectime = 0;
1781 tso->gran.fetchtime = 0;
1782 tso->gran.fetchcount = 0;
1783 tso->gran.blocktime = 0;
1784 tso->gran.blockcount = 0;
1785 tso->gran.blockedat = 0;
1786 tso->gran.globalsparks = 0;
1787 tso->gran.localsparks = 0;
1788 if (RtsFlags.GranFlags.Light)
1789 tso->gran.clock = Now; /* local clock */
1791 tso->gran.clock = 0;
1793 IF_DEBUG(gran,printTSO(tso));
1796 tso->par.magic = TSO_MAGIC; // debugging only
1798 tso->par.sparkname = 0;
1799 tso->par.startedat = CURRENT_TIME;
1800 tso->par.exported = 0;
1801 tso->par.basicblocks = 0;
1802 tso->par.allocs = 0;
1803 tso->par.exectime = 0;
1804 tso->par.fetchtime = 0;
1805 tso->par.fetchcount = 0;
1806 tso->par.blocktime = 0;
1807 tso->par.blockcount = 0;
1808 tso->par.blockedat = 0;
1809 tso->par.globalsparks = 0;
1810 tso->par.localsparks = 0;
1814 globalGranStats.tot_threads_created++;
1815 globalGranStats.threads_created_on_PE[CurrentProc]++;
1816 globalGranStats.tot_sq_len += spark_queue_len(CurrentProc);
1817 globalGranStats.tot_sq_probes++;
1819 // collect parallel global statistics (currently done together with GC stats)
1820 if (RtsFlags.ParFlags.ParStats.Global &&
1821 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
1822 //fprintf(stderr, "Creating thread %d @ %11.2f\n", tso->id, usertime());
1823 globalParStats.tot_threads_created++;
1829 belch("==__ schedule: Created TSO %d (%p);",
1830 CurrentProc, tso, tso->id));
1832 IF_PAR_DEBUG(verbose,
1833 belch("==__ schedule: Created TSO %d (%p); %d threads active",
1834 tso->id, tso, advisory_thread_count));
1836 IF_DEBUG(scheduler,sched_belch("created thread %ld, stack size = %lx words",
1837 tso->id, tso->stack_size));
1844 all parallel thread creation calls should fall through the following routine.
1847 createSparkThread(rtsSpark spark)
1849 ASSERT(spark != (rtsSpark)NULL);
1850 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads)
1852 barf("{createSparkThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1853 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1854 return END_TSO_QUEUE;
1858 tso = createThread(RtsFlags.GcFlags.initialStkSize);
1859 if (tso==END_TSO_QUEUE)
1860 barf("createSparkThread: Cannot create TSO");
1862 tso->priority = AdvisoryPriority;
1864 pushClosure(tso,spark);
1865 PUSH_ON_RUN_QUEUE(tso);
1866 advisory_thread_count++;
1873 Turn a spark into a thread.
1874 ToDo: fix for SMP (needs to acquire SCHED_MUTEX!)
1878 activateSpark (rtsSpark spark)
1882 tso = createSparkThread(spark);
1883 if (RtsFlags.ParFlags.ParStats.Full) {
1884 //ASSERT(run_queue_hd == END_TSO_QUEUE); // I think ...
1885 IF_PAR_DEBUG(verbose,
1886 belch("==^^ activateSpark: turning spark of closure %p (%s) into a thread",
1887 (StgClosure *)spark, info_type((StgClosure *)spark)));
1889 // ToDo: fwd info on local/global spark to thread -- HWL
1890 // tso->gran.exported = spark->exported;
1891 // tso->gran.locked = !spark->global;
1892 // tso->gran.sparkname = spark->name;
1898 static SchedulerStatus waitThread_(/*out*/StgMainThread* m,
1899 Capability *initialCapability
1903 /* ---------------------------------------------------------------------------
1906 * scheduleThread puts a thread on the head of the runnable queue.
1907 * This will usually be done immediately after a thread is created.
1908 * The caller of scheduleThread must create the thread using e.g.
1909 * createThread and push an appropriate closure
1910 * on this thread's stack before the scheduler is invoked.
1911 * ------------------------------------------------------------------------ */
1913 static void scheduleThread_ (StgTSO* tso);
1916 scheduleThread_(StgTSO *tso)
1918 // Precondition: sched_mutex must be held.
1919 PUSH_ON_RUN_QUEUE(tso);
1924 scheduleThread(StgTSO* tso)
1926 ACQUIRE_LOCK(&sched_mutex);
1927 scheduleThread_(tso);
1928 RELEASE_LOCK(&sched_mutex);
1931 #if defined(RTS_SUPPORTS_THREADS)
1932 static Condition bound_cond_cache;
1933 static int bound_cond_cache_full = 0;
1938 scheduleWaitThread(StgTSO* tso, /*[out]*/HaskellObj* ret,
1939 Capability *initialCapability)
1941 // Precondition: sched_mutex must be held
1944 m = stgMallocBytes(sizeof(StgMainThread), "waitThread");
1949 m->link = main_threads;
1951 if (main_threads != NULL) {
1952 main_threads->prev = m;
1956 #if defined(RTS_SUPPORTS_THREADS)
1957 // Allocating a new condition for each thread is expensive, so we
1958 // cache one. This is a pretty feeble hack, but it helps speed up
1959 // consecutive call-ins quite a bit.
1960 if (bound_cond_cache_full) {
1961 m->bound_thread_cond = bound_cond_cache;
1962 bound_cond_cache_full = 0;
1964 initCondition(&m->bound_thread_cond);
1968 /* Put the thread on the main-threads list prior to scheduling the TSO.
1969 Failure to do so introduces a race condition in the MT case (as
1970 identified by Wolfgang Thaller), whereby the new task/OS thread
1971 created by scheduleThread_() would complete prior to the thread
1972 that spawned it managed to put 'itself' on the main-threads list.
1973 The upshot of it all being that the worker thread wouldn't get to
1974 signal the completion of the its work item for the main thread to
1975 see (==> it got stuck waiting.) -- sof 6/02.
1977 IF_DEBUG(scheduler, sched_belch("waiting for thread (%d)", tso->id));
1979 PUSH_ON_RUN_QUEUE(tso);
1980 // NB. Don't call THREAD_RUNNABLE() here, because the thread is
1981 // bound and only runnable by *this* OS thread, so waking up other
1982 // workers will just slow things down.
1984 return waitThread_(m, initialCapability);
1987 /* ---------------------------------------------------------------------------
1990 * Initialise the scheduler. This resets all the queues - if the
1991 * queues contained any threads, they'll be garbage collected at the
1994 * ------------------------------------------------------------------------ */
2002 for (i=0; i<=MAX_PROC; i++) {
2003 run_queue_hds[i] = END_TSO_QUEUE;
2004 run_queue_tls[i] = END_TSO_QUEUE;
2005 blocked_queue_hds[i] = END_TSO_QUEUE;
2006 blocked_queue_tls[i] = END_TSO_QUEUE;
2007 ccalling_threadss[i] = END_TSO_QUEUE;
2008 sleeping_queue = END_TSO_QUEUE;
2011 run_queue_hd = END_TSO_QUEUE;
2012 run_queue_tl = END_TSO_QUEUE;
2013 blocked_queue_hd = END_TSO_QUEUE;
2014 blocked_queue_tl = END_TSO_QUEUE;
2015 sleeping_queue = END_TSO_QUEUE;
2018 suspended_ccalling_threads = END_TSO_QUEUE;
2020 main_threads = NULL;
2021 all_threads = END_TSO_QUEUE;
2026 RtsFlags.ConcFlags.ctxtSwitchTicks =
2027 RtsFlags.ConcFlags.ctxtSwitchTime / TICK_MILLISECS;
2029 #if defined(RTS_SUPPORTS_THREADS)
2030 /* Initialise the mutex and condition variables used by
2032 initMutex(&sched_mutex);
2033 initMutex(&term_mutex);
2036 ACQUIRE_LOCK(&sched_mutex);
2038 /* A capability holds the state a native thread needs in
2039 * order to execute STG code. At least one capability is
2040 * floating around (only SMP builds have more than one).
2044 #if defined(RTS_SUPPORTS_THREADS)
2045 /* start our haskell execution tasks */
2046 startTaskManager(0,taskStart);
2049 #if /* defined(SMP) ||*/ defined(PAR)
2053 RELEASE_LOCK(&sched_mutex);
2057 exitScheduler( void )
2059 #if defined(RTS_SUPPORTS_THREADS)
2062 shutting_down_scheduler = rtsTrue;
2065 /* ----------------------------------------------------------------------------
2066 Managing the per-task allocation areas.
2068 Each capability comes with an allocation area. These are
2069 fixed-length block lists into which allocation can be done.
2071 ToDo: no support for two-space collection at the moment???
2072 ------------------------------------------------------------------------- */
2076 waitThread_(StgMainThread* m, Capability *initialCapability)
2078 SchedulerStatus stat;
2080 // Precondition: sched_mutex must be held.
2081 IF_DEBUG(scheduler, sched_belch("new main thread (%d)", m->tso->id));
2084 /* GranSim specific init */
2085 CurrentTSO = m->tso; // the TSO to run
2086 procStatus[MainProc] = Busy; // status of main PE
2087 CurrentProc = MainProc; // PE to run it on
2088 schedule(m,initialCapability);
2090 schedule(m,initialCapability);
2091 ASSERT(m->stat != NoStatus);
2096 #if defined(RTS_SUPPORTS_THREADS)
2097 // Free the condition variable, returning it to the cache if possible.
2098 if (!bound_cond_cache_full) {
2099 bound_cond_cache = m->bound_thread_cond;
2100 bound_cond_cache_full = 1;
2102 closeCondition(&m->bound_thread_cond);
2106 IF_DEBUG(scheduler, sched_belch("main thread (%d) finished", m->tso->id));
2109 // Postcondition: sched_mutex still held
2113 /* ---------------------------------------------------------------------------
2114 Where are the roots that we know about?
2116 - all the threads on the runnable queue
2117 - all the threads on the blocked queue
2118 - all the threads on the sleeping queue
2119 - all the thread currently executing a _ccall_GC
2120 - all the "main threads"
2122 ------------------------------------------------------------------------ */
2124 /* This has to be protected either by the scheduler monitor, or by the
2125 garbage collection monitor (probably the latter).
2130 GetRoots( evac_fn evac )
2135 for (i=0; i<=RtsFlags.GranFlags.proc; i++) {
2136 if ((run_queue_hds[i] != END_TSO_QUEUE) && ((run_queue_hds[i] != NULL)))
2137 evac((StgClosure **)&run_queue_hds[i]);
2138 if ((run_queue_tls[i] != END_TSO_QUEUE) && ((run_queue_tls[i] != NULL)))
2139 evac((StgClosure **)&run_queue_tls[i]);
2141 if ((blocked_queue_hds[i] != END_TSO_QUEUE) && ((blocked_queue_hds[i] != NULL)))
2142 evac((StgClosure **)&blocked_queue_hds[i]);
2143 if ((blocked_queue_tls[i] != END_TSO_QUEUE) && ((blocked_queue_tls[i] != NULL)))
2144 evac((StgClosure **)&blocked_queue_tls[i]);
2145 if ((ccalling_threadss[i] != END_TSO_QUEUE) && ((ccalling_threadss[i] != NULL)))
2146 evac((StgClosure **)&ccalling_threads[i]);
2153 if (run_queue_hd != END_TSO_QUEUE) {
2154 ASSERT(run_queue_tl != END_TSO_QUEUE);
2155 evac((StgClosure **)&run_queue_hd);
2156 evac((StgClosure **)&run_queue_tl);
2159 if (blocked_queue_hd != END_TSO_QUEUE) {
2160 ASSERT(blocked_queue_tl != END_TSO_QUEUE);
2161 evac((StgClosure **)&blocked_queue_hd);
2162 evac((StgClosure **)&blocked_queue_tl);
2165 if (sleeping_queue != END_TSO_QUEUE) {
2166 evac((StgClosure **)&sleeping_queue);
2170 if (suspended_ccalling_threads != END_TSO_QUEUE) {
2171 evac((StgClosure **)&suspended_ccalling_threads);
2174 #if defined(PAR) || defined(GRAN)
2175 markSparkQueue(evac);
2178 #if defined(RTS_USER_SIGNALS)
2179 // mark the signal handlers (signals should be already blocked)
2180 markSignalHandlers(evac);
2184 /* -----------------------------------------------------------------------------
2187 This is the interface to the garbage collector from Haskell land.
2188 We provide this so that external C code can allocate and garbage
2189 collect when called from Haskell via _ccall_GC.
2191 It might be useful to provide an interface whereby the programmer
2192 can specify more roots (ToDo).
2194 This needs to be protected by the GC condition variable above. KH.
2195 -------------------------------------------------------------------------- */
2197 static void (*extra_roots)(evac_fn);
2202 /* Obligated to hold this lock upon entry */
2203 ACQUIRE_LOCK(&sched_mutex);
2204 GarbageCollect(GetRoots,rtsFalse);
2205 RELEASE_LOCK(&sched_mutex);
2209 performMajorGC(void)
2211 ACQUIRE_LOCK(&sched_mutex);
2212 GarbageCollect(GetRoots,rtsTrue);
2213 RELEASE_LOCK(&sched_mutex);
2217 AllRoots(evac_fn evac)
2219 GetRoots(evac); // the scheduler's roots
2220 extra_roots(evac); // the user's roots
2224 performGCWithRoots(void (*get_roots)(evac_fn))
2226 ACQUIRE_LOCK(&sched_mutex);
2227 extra_roots = get_roots;
2228 GarbageCollect(AllRoots,rtsFalse);
2229 RELEASE_LOCK(&sched_mutex);
2232 /* -----------------------------------------------------------------------------
2235 If the thread has reached its maximum stack size, then raise the
2236 StackOverflow exception in the offending thread. Otherwise
2237 relocate the TSO into a larger chunk of memory and adjust its stack
2239 -------------------------------------------------------------------------- */
2242 threadStackOverflow(StgTSO *tso)
2244 nat new_stack_size, new_tso_size, stack_words;
2248 IF_DEBUG(sanity,checkTSO(tso));
2249 if (tso->stack_size >= tso->max_stack_size) {
2252 belch("@@ threadStackOverflow of TSO %d (%p): stack too large (now %ld; max is %ld)",
2253 tso->id, tso, tso->stack_size, tso->max_stack_size);
2254 /* If we're debugging, just print out the top of the stack */
2255 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2258 /* Send this thread the StackOverflow exception */
2259 raiseAsync(tso, (StgClosure *)stackOverflow_closure);
2263 /* Try to double the current stack size. If that takes us over the
2264 * maximum stack size for this thread, then use the maximum instead.
2265 * Finally round up so the TSO ends up as a whole number of blocks.
2267 new_stack_size = stg_min(tso->stack_size * 2, tso->max_stack_size);
2268 new_tso_size = (nat)BLOCK_ROUND_UP(new_stack_size * sizeof(W_) +
2269 TSO_STRUCT_SIZE)/sizeof(W_);
2270 new_tso_size = round_to_mblocks(new_tso_size); /* Be MBLOCK-friendly */
2271 new_stack_size = new_tso_size - TSO_STRUCT_SIZEW;
2273 IF_DEBUG(scheduler, fprintf(stderr,"== sched: increasing stack size from %d words to %d.\n", tso->stack_size, new_stack_size));
2275 dest = (StgTSO *)allocate(new_tso_size);
2276 TICK_ALLOC_TSO(new_stack_size,0);
2278 /* copy the TSO block and the old stack into the new area */
2279 memcpy(dest,tso,TSO_STRUCT_SIZE);
2280 stack_words = tso->stack + tso->stack_size - tso->sp;
2281 new_sp = (P_)dest + new_tso_size - stack_words;
2282 memcpy(new_sp, tso->sp, stack_words * sizeof(W_));
2284 /* relocate the stack pointers... */
2286 dest->stack_size = new_stack_size;
2288 /* Mark the old TSO as relocated. We have to check for relocated
2289 * TSOs in the garbage collector and any primops that deal with TSOs.
2291 * It's important to set the sp value to just beyond the end
2292 * of the stack, so we don't attempt to scavenge any part of the
2295 tso->what_next = ThreadRelocated;
2297 tso->sp = (P_)&(tso->stack[tso->stack_size]);
2298 tso->why_blocked = NotBlocked;
2299 dest->mut_link = NULL;
2301 IF_PAR_DEBUG(verbose,
2302 belch("@@ threadStackOverflow of TSO %d (now at %p): stack size increased to %ld",
2303 tso->id, tso, tso->stack_size);
2304 /* If we're debugging, just print out the top of the stack */
2305 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2308 IF_DEBUG(sanity,checkTSO(tso));
2310 IF_DEBUG(scheduler,printTSO(dest));
2316 /* ---------------------------------------------------------------------------
2317 Wake up a queue that was blocked on some resource.
2318 ------------------------------------------------------------------------ */
2322 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2327 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2329 /* write RESUME events to log file and
2330 update blocked and fetch time (depending on type of the orig closure) */
2331 if (RtsFlags.ParFlags.ParStats.Full) {
2332 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
2333 GR_RESUMEQ, ((StgTSO *)bqe), ((StgTSO *)bqe)->block_info.closure,
2334 0, 0 /* spark_queue_len(ADVISORY_POOL) */);
2335 if (EMPTY_RUN_QUEUE())
2336 emitSchedule = rtsTrue;
2338 switch (get_itbl(node)->type) {
2340 ((StgTSO *)bqe)->par.fetchtime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2345 ((StgTSO *)bqe)->par.blocktime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2352 barf("{unblockOneLocked}Daq Qagh: unexpected closure in blocking queue");
2359 static StgBlockingQueueElement *
2360 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2363 PEs node_loc, tso_loc;
2365 node_loc = where_is(node); // should be lifted out of loop
2366 tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2367 tso_loc = where_is((StgClosure *)tso);
2368 if (IS_LOCAL_TO(PROCS(node),tso_loc)) { // TSO is local
2369 /* !fake_fetch => TSO is on CurrentProc is same as IS_LOCAL_TO */
2370 ASSERT(CurrentProc!=node_loc || tso_loc==CurrentProc);
2371 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.lunblocktime;
2372 // insertThread(tso, node_loc);
2373 new_event(tso_loc, tso_loc, CurrentTime[CurrentProc],
2375 tso, node, (rtsSpark*)NULL);
2376 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2379 } else { // TSO is remote (actually should be FMBQ)
2380 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.mpacktime +
2381 RtsFlags.GranFlags.Costs.gunblocktime +
2382 RtsFlags.GranFlags.Costs.latency;
2383 new_event(tso_loc, CurrentProc, CurrentTime[CurrentProc],
2385 tso, node, (rtsSpark*)NULL);
2386 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2389 /* the thread-queue-overhead is accounted for in either Resume or UnblockThread */
2391 fprintf(stderr," %s TSO %d (%p) [PE %d] (block_info.closure=%p) (next=%p) ,",
2392 (node_loc==tso_loc ? "Local" : "Global"),
2393 tso->id, tso, CurrentProc, tso->block_info.closure, tso->link));
2394 tso->block_info.closure = NULL;
2395 IF_DEBUG(scheduler,belch("-- Waking up thread %ld (%p)",
2399 static StgBlockingQueueElement *
2400 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2402 StgBlockingQueueElement *next;
2404 switch (get_itbl(bqe)->type) {
2406 ASSERT(((StgTSO *)bqe)->why_blocked != NotBlocked);
2407 /* if it's a TSO just push it onto the run_queue */
2409 // ((StgTSO *)bqe)->link = END_TSO_QUEUE; // debugging?
2410 PUSH_ON_RUN_QUEUE((StgTSO *)bqe);
2412 unblockCount(bqe, node);
2413 /* reset blocking status after dumping event */
2414 ((StgTSO *)bqe)->why_blocked = NotBlocked;
2418 /* if it's a BLOCKED_FETCH put it on the PendingFetches list */
2420 bqe->link = (StgBlockingQueueElement *)PendingFetches;
2421 PendingFetches = (StgBlockedFetch *)bqe;
2425 /* can ignore this case in a non-debugging setup;
2426 see comments on RBHSave closures above */
2428 /* check that the closure is an RBHSave closure */
2429 ASSERT(get_itbl((StgClosure *)bqe) == &stg_RBH_Save_0_info ||
2430 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_1_info ||
2431 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_2_info);
2435 barf("{unblockOneLocked}Daq Qagh: Unexpected IP (%#lx; %s) in blocking queue at %#lx\n",
2436 get_itbl((StgClosure *)bqe), info_type((StgClosure *)bqe),
2440 IF_PAR_DEBUG(bq, fprintf(stderr, ", %p (%s)", bqe, info_type((StgClosure*)bqe)));
2444 #else /* !GRAN && !PAR */
2446 unblockOneLocked(StgTSO *tso)
2450 ASSERT(get_itbl(tso)->type == TSO);
2451 ASSERT(tso->why_blocked != NotBlocked);
2452 tso->why_blocked = NotBlocked;
2454 PUSH_ON_RUN_QUEUE(tso);
2456 IF_DEBUG(scheduler,sched_belch("waking up thread %ld", tso->id));
2461 #if defined(GRAN) || defined(PAR)
2462 INLINE_ME StgBlockingQueueElement *
2463 unblockOne(StgBlockingQueueElement *bqe, StgClosure *node)
2465 ACQUIRE_LOCK(&sched_mutex);
2466 bqe = unblockOneLocked(bqe, node);
2467 RELEASE_LOCK(&sched_mutex);
2472 unblockOne(StgTSO *tso)
2474 ACQUIRE_LOCK(&sched_mutex);
2475 tso = unblockOneLocked(tso);
2476 RELEASE_LOCK(&sched_mutex);
2483 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2485 StgBlockingQueueElement *bqe;
2490 belch("##-_ AwBQ for node %p on PE %d @ %ld by TSO %d (%p): ", \
2491 node, CurrentProc, CurrentTime[CurrentProc],
2492 CurrentTSO->id, CurrentTSO));
2494 node_loc = where_is(node);
2496 ASSERT(q == END_BQ_QUEUE ||
2497 get_itbl(q)->type == TSO || // q is either a TSO or an RBHSave
2498 get_itbl(q)->type == CONSTR); // closure (type constructor)
2499 ASSERT(is_unique(node));
2501 /* FAKE FETCH: magically copy the node to the tso's proc;
2502 no Fetch necessary because in reality the node should not have been
2503 moved to the other PE in the first place
2505 if (CurrentProc!=node_loc) {
2507 belch("## node %p is on PE %d but CurrentProc is %d (TSO %d); assuming fake fetch and adjusting bitmask (old: %#x)",
2508 node, node_loc, CurrentProc, CurrentTSO->id,
2509 // CurrentTSO, where_is(CurrentTSO),
2510 node->header.gran.procs));
2511 node->header.gran.procs = (node->header.gran.procs) | PE_NUMBER(CurrentProc);
2513 belch("## new bitmask of node %p is %#x",
2514 node, node->header.gran.procs));
2515 if (RtsFlags.GranFlags.GranSimStats.Global) {
2516 globalGranStats.tot_fake_fetches++;
2521 // ToDo: check: ASSERT(CurrentProc==node_loc);
2522 while (get_itbl(bqe)->type==TSO) { // q != END_TSO_QUEUE) {
2525 bqe points to the current element in the queue
2526 next points to the next element in the queue
2528 //tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2529 //tso_loc = where_is(tso);
2531 bqe = unblockOneLocked(bqe, node);
2534 /* if this is the BQ of an RBH, we have to put back the info ripped out of
2535 the closure to make room for the anchor of the BQ */
2536 if (bqe!=END_BQ_QUEUE) {
2537 ASSERT(get_itbl(node)->type == RBH && get_itbl(bqe)->type == CONSTR);
2539 ASSERT((info_ptr==&RBH_Save_0_info) ||
2540 (info_ptr==&RBH_Save_1_info) ||
2541 (info_ptr==&RBH_Save_2_info));
2543 /* cf. convertToRBH in RBH.c for writing the RBHSave closure */
2544 ((StgRBH *)node)->blocking_queue = (StgBlockingQueueElement *)((StgRBHSave *)bqe)->payload[0];
2545 ((StgRBH *)node)->mut_link = (StgMutClosure *)((StgRBHSave *)bqe)->payload[1];
2548 belch("## Filled in RBH_Save for %p (%s) at end of AwBQ",
2549 node, info_type(node)));
2552 /* statistics gathering */
2553 if (RtsFlags.GranFlags.GranSimStats.Global) {
2554 // globalGranStats.tot_bq_processing_time += bq_processing_time;
2555 globalGranStats.tot_bq_len += len; // total length of all bqs awakened
2556 // globalGranStats.tot_bq_len_local += len_local; // same for local TSOs only
2557 globalGranStats.tot_awbq++; // total no. of bqs awakened
2560 fprintf(stderr,"## BQ Stats of %p: [%d entries] %s\n",
2561 node, len, (bqe!=END_BQ_QUEUE) ? "RBH" : ""));
2565 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2567 StgBlockingQueueElement *bqe;
2569 ACQUIRE_LOCK(&sched_mutex);
2571 IF_PAR_DEBUG(verbose,
2572 belch("##-_ AwBQ for node %p on [%x]: ",
2576 if(get_itbl(q)->type == CONSTR || q==END_BQ_QUEUE) {
2577 IF_PAR_DEBUG(verbose, belch("## ... nothing to unblock so lets just return. RFP (BUG?)"));
2582 ASSERT(q == END_BQ_QUEUE ||
2583 get_itbl(q)->type == TSO ||
2584 get_itbl(q)->type == BLOCKED_FETCH ||
2585 get_itbl(q)->type == CONSTR);
2588 while (get_itbl(bqe)->type==TSO ||
2589 get_itbl(bqe)->type==BLOCKED_FETCH) {
2590 bqe = unblockOneLocked(bqe, node);
2592 RELEASE_LOCK(&sched_mutex);
2595 #else /* !GRAN && !PAR */
2598 awakenBlockedQueueNoLock(StgTSO *tso)
2600 while (tso != END_TSO_QUEUE) {
2601 tso = unblockOneLocked(tso);
2606 awakenBlockedQueue(StgTSO *tso)
2608 ACQUIRE_LOCK(&sched_mutex);
2609 while (tso != END_TSO_QUEUE) {
2610 tso = unblockOneLocked(tso);
2612 RELEASE_LOCK(&sched_mutex);
2616 /* ---------------------------------------------------------------------------
2618 - usually called inside a signal handler so it mustn't do anything fancy.
2619 ------------------------------------------------------------------------ */
2622 interruptStgRts(void)
2626 #ifdef RTS_SUPPORTS_THREADS
2627 wakeBlockedWorkerThread();
2631 /* -----------------------------------------------------------------------------
2634 This is for use when we raise an exception in another thread, which
2636 This has nothing to do with the UnblockThread event in GranSim. -- HWL
2637 -------------------------------------------------------------------------- */
2639 #if defined(GRAN) || defined(PAR)
2641 NB: only the type of the blocking queue is different in GranSim and GUM
2642 the operations on the queue-elements are the same
2643 long live polymorphism!
2645 Locks: sched_mutex is held upon entry and exit.
2649 unblockThread(StgTSO *tso)
2651 StgBlockingQueueElement *t, **last;
2653 switch (tso->why_blocked) {
2656 return; /* not blocked */
2659 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
2661 StgBlockingQueueElement *last_tso = END_BQ_QUEUE;
2662 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
2664 last = (StgBlockingQueueElement **)&mvar->head;
2665 for (t = (StgBlockingQueueElement *)mvar->head;
2667 last = &t->link, last_tso = t, t = t->link) {
2668 if (t == (StgBlockingQueueElement *)tso) {
2669 *last = (StgBlockingQueueElement *)tso->link;
2670 if (mvar->tail == tso) {
2671 mvar->tail = (StgTSO *)last_tso;
2676 barf("unblockThread (MVAR): TSO not found");
2679 case BlockedOnBlackHole:
2680 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
2682 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
2684 last = &bq->blocking_queue;
2685 for (t = bq->blocking_queue;
2687 last = &t->link, t = t->link) {
2688 if (t == (StgBlockingQueueElement *)tso) {
2689 *last = (StgBlockingQueueElement *)tso->link;
2693 barf("unblockThread (BLACKHOLE): TSO not found");
2696 case BlockedOnException:
2698 StgTSO *target = tso->block_info.tso;
2700 ASSERT(get_itbl(target)->type == TSO);
2702 if (target->what_next == ThreadRelocated) {
2703 target = target->link;
2704 ASSERT(get_itbl(target)->type == TSO);
2707 ASSERT(target->blocked_exceptions != NULL);
2709 last = (StgBlockingQueueElement **)&target->blocked_exceptions;
2710 for (t = (StgBlockingQueueElement *)target->blocked_exceptions;
2712 last = &t->link, t = t->link) {
2713 ASSERT(get_itbl(t)->type == TSO);
2714 if (t == (StgBlockingQueueElement *)tso) {
2715 *last = (StgBlockingQueueElement *)tso->link;
2719 barf("unblockThread (Exception): TSO not found");
2723 case BlockedOnWrite:
2724 #if defined(mingw32_TARGET_OS)
2725 case BlockedOnDoProc:
2728 /* take TSO off blocked_queue */
2729 StgBlockingQueueElement *prev = NULL;
2730 for (t = (StgBlockingQueueElement *)blocked_queue_hd; t != END_BQ_QUEUE;
2731 prev = t, t = t->link) {
2732 if (t == (StgBlockingQueueElement *)tso) {
2734 blocked_queue_hd = (StgTSO *)t->link;
2735 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
2736 blocked_queue_tl = END_TSO_QUEUE;
2739 prev->link = t->link;
2740 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
2741 blocked_queue_tl = (StgTSO *)prev;
2747 barf("unblockThread (I/O): TSO not found");
2750 case BlockedOnDelay:
2752 /* take TSO off sleeping_queue */
2753 StgBlockingQueueElement *prev = NULL;
2754 for (t = (StgBlockingQueueElement *)sleeping_queue; t != END_BQ_QUEUE;
2755 prev = t, t = t->link) {
2756 if (t == (StgBlockingQueueElement *)tso) {
2758 sleeping_queue = (StgTSO *)t->link;
2760 prev->link = t->link;
2765 barf("unblockThread (delay): TSO not found");
2769 barf("unblockThread");
2773 tso->link = END_TSO_QUEUE;
2774 tso->why_blocked = NotBlocked;
2775 tso->block_info.closure = NULL;
2776 PUSH_ON_RUN_QUEUE(tso);
2780 unblockThread(StgTSO *tso)
2784 /* To avoid locking unnecessarily. */
2785 if (tso->why_blocked == NotBlocked) {
2789 switch (tso->why_blocked) {
2792 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
2794 StgTSO *last_tso = END_TSO_QUEUE;
2795 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
2798 for (t = mvar->head; t != END_TSO_QUEUE;
2799 last = &t->link, last_tso = t, t = t->link) {
2802 if (mvar->tail == tso) {
2803 mvar->tail = last_tso;
2808 barf("unblockThread (MVAR): TSO not found");
2811 case BlockedOnBlackHole:
2812 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
2814 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
2816 last = &bq->blocking_queue;
2817 for (t = bq->blocking_queue; t != END_TSO_QUEUE;
2818 last = &t->link, t = t->link) {
2824 barf("unblockThread (BLACKHOLE): TSO not found");
2827 case BlockedOnException:
2829 StgTSO *target = tso->block_info.tso;
2831 ASSERT(get_itbl(target)->type == TSO);
2833 while (target->what_next == ThreadRelocated) {
2834 target = target->link;
2835 ASSERT(get_itbl(target)->type == TSO);
2838 ASSERT(target->blocked_exceptions != NULL);
2840 last = &target->blocked_exceptions;
2841 for (t = target->blocked_exceptions; t != END_TSO_QUEUE;
2842 last = &t->link, t = t->link) {
2843 ASSERT(get_itbl(t)->type == TSO);
2849 barf("unblockThread (Exception): TSO not found");
2853 case BlockedOnWrite:
2854 #if defined(mingw32_TARGET_OS)
2855 case BlockedOnDoProc:
2858 StgTSO *prev = NULL;
2859 for (t = blocked_queue_hd; t != END_TSO_QUEUE;
2860 prev = t, t = t->link) {
2863 blocked_queue_hd = t->link;
2864 if (blocked_queue_tl == t) {
2865 blocked_queue_tl = END_TSO_QUEUE;
2868 prev->link = t->link;
2869 if (blocked_queue_tl == t) {
2870 blocked_queue_tl = prev;
2876 barf("unblockThread (I/O): TSO not found");
2879 case BlockedOnDelay:
2881 StgTSO *prev = NULL;
2882 for (t = sleeping_queue; t != END_TSO_QUEUE;
2883 prev = t, t = t->link) {
2886 sleeping_queue = t->link;
2888 prev->link = t->link;
2893 barf("unblockThread (delay): TSO not found");
2897 barf("unblockThread");
2901 tso->link = END_TSO_QUEUE;
2902 tso->why_blocked = NotBlocked;
2903 tso->block_info.closure = NULL;
2904 PUSH_ON_RUN_QUEUE(tso);
2908 /* -----------------------------------------------------------------------------
2911 * The following function implements the magic for raising an
2912 * asynchronous exception in an existing thread.
2914 * We first remove the thread from any queue on which it might be
2915 * blocked. The possible blockages are MVARs and BLACKHOLE_BQs.
2917 * We strip the stack down to the innermost CATCH_FRAME, building
2918 * thunks in the heap for all the active computations, so they can
2919 * be restarted if necessary. When we reach a CATCH_FRAME, we build
2920 * an application of the handler to the exception, and push it on
2921 * the top of the stack.
2923 * How exactly do we save all the active computations? We create an
2924 * AP_STACK for every UpdateFrame on the stack. Entering one of these
2925 * AP_STACKs pushes everything from the corresponding update frame
2926 * upwards onto the stack. (Actually, it pushes everything up to the
2927 * next update frame plus a pointer to the next AP_STACK object.
2928 * Entering the next AP_STACK object pushes more onto the stack until we
2929 * reach the last AP_STACK object - at which point the stack should look
2930 * exactly as it did when we killed the TSO and we can continue
2931 * execution by entering the closure on top of the stack.
2933 * We can also kill a thread entirely - this happens if either (a) the
2934 * exception passed to raiseAsync is NULL, or (b) there's no
2935 * CATCH_FRAME on the stack. In either case, we strip the entire
2936 * stack and replace the thread with a zombie.
2938 * Locks: sched_mutex held upon entry nor exit.
2940 * -------------------------------------------------------------------------- */
2943 deleteThread(StgTSO *tso)
2945 raiseAsync(tso,NULL);
2948 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
2950 deleteThreadImmediately(StgTSO *tso)
2951 { // for forkProcess only:
2952 // delete thread without giving it a chance to catch the KillThread exception
2954 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
2958 if (tso->why_blocked != BlockedOnCCall &&
2959 tso->why_blocked != BlockedOnCCall_NoUnblockExc) {
2963 tso->what_next = ThreadKilled;
2968 raiseAsyncWithLock(StgTSO *tso, StgClosure *exception)
2970 /* When raising async exs from contexts where sched_mutex isn't held;
2971 use raiseAsyncWithLock(). */
2972 ACQUIRE_LOCK(&sched_mutex);
2973 raiseAsync(tso,exception);
2974 RELEASE_LOCK(&sched_mutex);
2978 raiseAsync(StgTSO *tso, StgClosure *exception)
2980 StgRetInfoTable *info;
2983 // Thread already dead?
2984 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
2989 sched_belch("raising exception in thread %ld.", tso->id));
2991 // Remove it from any blocking queues
2996 // The stack freezing code assumes there's a closure pointer on
2997 // the top of the stack, so we have to arrange that this is the case...
2999 if (sp[0] == (W_)&stg_enter_info) {
3003 sp[0] = (W_)&stg_dummy_ret_closure;
3009 // 1. Let the top of the stack be the "current closure"
3011 // 2. Walk up the stack until we find either an UPDATE_FRAME or a
3014 // 3. If it's an UPDATE_FRAME, then make an AP_STACK containing the
3015 // current closure applied to the chunk of stack up to (but not
3016 // including) the update frame. This closure becomes the "current
3017 // closure". Go back to step 2.
3019 // 4. If it's a CATCH_FRAME, then leave the exception handler on
3020 // top of the stack applied to the exception.
3022 // 5. If it's a STOP_FRAME, then kill the thread.
3027 info = get_ret_itbl((StgClosure *)frame);
3029 while (info->i.type != UPDATE_FRAME
3030 && (info->i.type != CATCH_FRAME || exception == NULL)
3031 && info->i.type != STOP_FRAME) {
3032 frame += stack_frame_sizeW((StgClosure *)frame);
3033 info = get_ret_itbl((StgClosure *)frame);
3036 switch (info->i.type) {
3039 // If we find a CATCH_FRAME, and we've got an exception to raise,
3040 // then build the THUNK raise(exception), and leave it on
3041 // top of the CATCH_FRAME ready to enter.
3045 StgCatchFrame *cf = (StgCatchFrame *)frame;
3049 // we've got an exception to raise, so let's pass it to the
3050 // handler in this frame.
3052 raise = (StgClosure *)allocate(sizeofW(StgClosure)+1);
3053 TICK_ALLOC_SE_THK(1,0);
3054 SET_HDR(raise,&stg_raise_info,cf->header.prof.ccs);
3055 raise->payload[0] = exception;
3057 // throw away the stack from Sp up to the CATCH_FRAME.
3061 /* Ensure that async excpetions are blocked now, so we don't get
3062 * a surprise exception before we get around to executing the
3065 if (tso->blocked_exceptions == NULL) {
3066 tso->blocked_exceptions = END_TSO_QUEUE;
3069 /* Put the newly-built THUNK on top of the stack, ready to execute
3070 * when the thread restarts.
3073 sp[-1] = (W_)&stg_enter_info;
3075 tso->what_next = ThreadRunGHC;
3076 IF_DEBUG(sanity, checkTSO(tso));
3085 // First build an AP_STACK consisting of the stack chunk above the
3086 // current update frame, with the top word on the stack as the
3089 words = frame - sp - 1;
3090 ap = (StgAP_STACK *)allocate(PAP_sizeW(words));
3093 ap->fun = (StgClosure *)sp[0];
3095 for(i=0; i < (nat)words; ++i) {
3096 ap->payload[i] = (StgClosure *)*sp++;
3099 SET_HDR(ap,&stg_AP_STACK_info,
3100 ((StgClosure *)frame)->header.prof.ccs /* ToDo */);
3101 TICK_ALLOC_UP_THK(words+1,0);
3104 fprintf(stderr, "sched: Updating ");
3105 printPtr((P_)((StgUpdateFrame *)frame)->updatee);
3106 fprintf(stderr, " with ");
3107 printObj((StgClosure *)ap);
3110 // Replace the updatee with an indirection - happily
3111 // this will also wake up any threads currently
3112 // waiting on the result.
3114 // Warning: if we're in a loop, more than one update frame on
3115 // the stack may point to the same object. Be careful not to
3116 // overwrite an IND_OLDGEN in this case, because we'll screw
3117 // up the mutable lists. To be on the safe side, don't
3118 // overwrite any kind of indirection at all. See also
3119 // threadSqueezeStack in GC.c, where we have to make a similar
3122 if (!closure_IND(((StgUpdateFrame *)frame)->updatee)) {
3123 // revert the black hole
3124 UPD_IND_NOLOCK(((StgUpdateFrame *)frame)->updatee,ap);
3126 sp += sizeofW(StgUpdateFrame) - 1;
3127 sp[0] = (W_)ap; // push onto stack
3132 // We've stripped the entire stack, the thread is now dead.
3133 sp += sizeofW(StgStopFrame);
3134 tso->what_next = ThreadKilled;
3145 /* -----------------------------------------------------------------------------
3146 resurrectThreads is called after garbage collection on the list of
3147 threads found to be garbage. Each of these threads will be woken
3148 up and sent a signal: BlockedOnDeadMVar if the thread was blocked
3149 on an MVar, or NonTermination if the thread was blocked on a Black
3152 Locks: sched_mutex isn't held upon entry nor exit.
3153 -------------------------------------------------------------------------- */
3156 resurrectThreads( StgTSO *threads )
3160 for (tso = threads; tso != END_TSO_QUEUE; tso = next) {
3161 next = tso->global_link;
3162 tso->global_link = all_threads;
3164 IF_DEBUG(scheduler, sched_belch("resurrecting thread %d", tso->id));
3166 switch (tso->why_blocked) {
3168 case BlockedOnException:
3169 /* Called by GC - sched_mutex lock is currently held. */
3170 raiseAsync(tso,(StgClosure *)BlockedOnDeadMVar_closure);
3172 case BlockedOnBlackHole:
3173 raiseAsync(tso,(StgClosure *)NonTermination_closure);
3176 /* This might happen if the thread was blocked on a black hole
3177 * belonging to a thread that we've just woken up (raiseAsync
3178 * can wake up threads, remember...).
3182 barf("resurrectThreads: thread blocked in a strange way");
3187 /* -----------------------------------------------------------------------------
3188 * Blackhole detection: if we reach a deadlock, test whether any
3189 * threads are blocked on themselves. Any threads which are found to
3190 * be self-blocked get sent a NonTermination exception.
3192 * This is only done in a deadlock situation in order to avoid
3193 * performance overhead in the normal case.
3195 * Locks: sched_mutex is held upon entry and exit.
3196 * -------------------------------------------------------------------------- */
3199 detectBlackHoles( void )
3201 StgTSO *tso = all_threads;
3203 StgClosure *blocked_on;
3204 StgRetInfoTable *info;
3206 for (tso = all_threads; tso != END_TSO_QUEUE; tso = tso->global_link) {
3208 while (tso->what_next == ThreadRelocated) {
3210 ASSERT(get_itbl(tso)->type == TSO);
3213 if (tso->why_blocked != BlockedOnBlackHole) {
3216 blocked_on = tso->block_info.closure;
3218 frame = (StgClosure *)tso->sp;
3221 info = get_ret_itbl(frame);
3222 switch (info->i.type) {
3224 if (((StgUpdateFrame *)frame)->updatee == blocked_on) {
3225 /* We are blocking on one of our own computations, so
3226 * send this thread the NonTermination exception.
3229 sched_belch("thread %d is blocked on itself", tso->id));
3230 raiseAsync(tso, (StgClosure *)NonTermination_closure);
3234 frame = (StgClosure *) ((StgUpdateFrame *)frame + 1);
3240 // normal stack frames; do nothing except advance the pointer
3242 (StgPtr)frame += stack_frame_sizeW(frame);
3249 /* ----------------------------------------------------------------------------
3250 * Debugging: why is a thread blocked
3251 * [Also provides useful information when debugging threaded programs
3252 * at the Haskell source code level, so enable outside of DEBUG. --sof 7/02]
3253 ------------------------------------------------------------------------- */
3257 printThreadBlockage(StgTSO *tso)
3259 switch (tso->why_blocked) {
3261 fprintf(stderr,"is blocked on read from fd %d", tso->block_info.fd);
3263 case BlockedOnWrite:
3264 fprintf(stderr,"is blocked on write to fd %d", tso->block_info.fd);
3266 #if defined(mingw32_TARGET_OS)
3267 case BlockedOnDoProc:
3268 fprintf(stderr,"is blocked on proc (request: %d)", tso->block_info.async_result->reqID);
3271 case BlockedOnDelay:
3272 fprintf(stderr,"is blocked until %d", tso->block_info.target);
3275 fprintf(stderr,"is blocked on an MVar");
3277 case BlockedOnException:
3278 fprintf(stderr,"is blocked on delivering an exception to thread %d",
3279 tso->block_info.tso->id);
3281 case BlockedOnBlackHole:
3282 fprintf(stderr,"is blocked on a black hole");
3285 fprintf(stderr,"is not blocked");
3289 fprintf(stderr,"is blocked on global address; local FM_BQ is %p (%s)",
3290 tso->block_info.closure, info_type(tso->block_info.closure));
3292 case BlockedOnGA_NoSend:
3293 fprintf(stderr,"is blocked on global address (no send); local FM_BQ is %p (%s)",
3294 tso->block_info.closure, info_type(tso->block_info.closure));
3297 case BlockedOnCCall:
3298 fprintf(stderr,"is blocked on an external call");
3300 case BlockedOnCCall_NoUnblockExc:
3301 fprintf(stderr,"is blocked on an external call (exceptions were already blocked)");
3304 barf("printThreadBlockage: strange tso->why_blocked: %d for TSO %d (%d)",
3305 tso->why_blocked, tso->id, tso);
3311 printThreadStatus(StgTSO *tso)
3313 switch (tso->what_next) {
3315 fprintf(stderr,"has been killed");
3317 case ThreadComplete:
3318 fprintf(stderr,"has completed");
3321 printThreadBlockage(tso);
3326 printAllThreads(void)
3332 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3333 ullong_format_string(TIME_ON_PROC(CurrentProc),
3334 time_string, rtsFalse/*no commas!*/);
3336 fprintf(stderr, "all threads at [%s]:\n", time_string);
3338 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3339 ullong_format_string(CURRENT_TIME,
3340 time_string, rtsFalse/*no commas!*/);
3342 fprintf(stderr,"all threads at [%s]:\n", time_string);
3344 fprintf(stderr,"all threads:\n");
3347 for (t = all_threads; t != END_TSO_QUEUE; t = t->global_link) {
3348 fprintf(stderr, "\tthread %d @ %p ", t->id, (void *)t);
3349 label = lookupThreadLabel(t->id);
3350 if (label) fprintf(stderr,"[\"%s\"] ",(char *)label);
3351 printThreadStatus(t);
3352 fprintf(stderr,"\n");
3359 Print a whole blocking queue attached to node (debugging only).
3363 print_bq (StgClosure *node)
3365 StgBlockingQueueElement *bqe;
3369 fprintf(stderr,"## BQ of closure %p (%s): ",
3370 node, info_type(node));
3372 /* should cover all closures that may have a blocking queue */
3373 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3374 get_itbl(node)->type == FETCH_ME_BQ ||
3375 get_itbl(node)->type == RBH ||
3376 get_itbl(node)->type == MVAR);
3378 ASSERT(node!=(StgClosure*)NULL); // sanity check
3380 print_bqe(((StgBlockingQueue*)node)->blocking_queue);
3384 Print a whole blocking queue starting with the element bqe.
3387 print_bqe (StgBlockingQueueElement *bqe)
3392 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3394 for (end = (bqe==END_BQ_QUEUE);
3395 !end; // iterate until bqe points to a CONSTR
3396 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE),
3397 bqe = end ? END_BQ_QUEUE : bqe->link) {
3398 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3399 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3400 /* types of closures that may appear in a blocking queue */
3401 ASSERT(get_itbl(bqe)->type == TSO ||
3402 get_itbl(bqe)->type == BLOCKED_FETCH ||
3403 get_itbl(bqe)->type == CONSTR);
3404 /* only BQs of an RBH end with an RBH_Save closure */
3405 //ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3407 switch (get_itbl(bqe)->type) {
3409 fprintf(stderr," TSO %u (%x),",
3410 ((StgTSO *)bqe)->id, ((StgTSO *)bqe));
3413 fprintf(stderr," BF (node=%p, ga=((%x, %d, %x)),",
3414 ((StgBlockedFetch *)bqe)->node,
3415 ((StgBlockedFetch *)bqe)->ga.payload.gc.gtid,
3416 ((StgBlockedFetch *)bqe)->ga.payload.gc.slot,
3417 ((StgBlockedFetch *)bqe)->ga.weight);
3420 fprintf(stderr," %s (IP %p),",
3421 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3422 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3423 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3424 "RBH_Save_?"), get_itbl(bqe));
3427 barf("Unexpected closure type %s in blocking queue", // of %p (%s)",
3428 info_type((StgClosure *)bqe)); // , node, info_type(node));
3432 fputc('\n', stderr);
3434 # elif defined(GRAN)
3436 print_bq (StgClosure *node)
3438 StgBlockingQueueElement *bqe;
3439 PEs node_loc, tso_loc;
3442 /* should cover all closures that may have a blocking queue */
3443 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3444 get_itbl(node)->type == FETCH_ME_BQ ||
3445 get_itbl(node)->type == RBH);
3447 ASSERT(node!=(StgClosure*)NULL); // sanity check
3448 node_loc = where_is(node);
3450 fprintf(stderr,"## BQ of closure %p (%s) on [PE %d]: ",
3451 node, info_type(node), node_loc);
3454 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3456 for (bqe = ((StgBlockingQueue*)node)->blocking_queue, end = (bqe==END_BQ_QUEUE);
3457 !end; // iterate until bqe points to a CONSTR
3458 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE), bqe = end ? END_BQ_QUEUE : bqe->link) {
3459 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3460 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3461 /* types of closures that may appear in a blocking queue */
3462 ASSERT(get_itbl(bqe)->type == TSO ||
3463 get_itbl(bqe)->type == CONSTR);
3464 /* only BQs of an RBH end with an RBH_Save closure */
3465 ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3467 tso_loc = where_is((StgClosure *)bqe);
3468 switch (get_itbl(bqe)->type) {
3470 fprintf(stderr," TSO %d (%p) on [PE %d],",
3471 ((StgTSO *)bqe)->id, (StgTSO *)bqe, tso_loc);
3474 fprintf(stderr," %s (IP %p),",
3475 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3476 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3477 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3478 "RBH_Save_?"), get_itbl(bqe));
3481 barf("Unexpected closure type %s in blocking queue of %p (%s)",
3482 info_type((StgClosure *)bqe), node, info_type(node));
3486 fputc('\n', stderr);
3490 Nice and easy: only TSOs on the blocking queue
3493 print_bq (StgClosure *node)
3497 ASSERT(node!=(StgClosure*)NULL); // sanity check
3498 for (tso = ((StgBlockingQueue*)node)->blocking_queue;
3499 tso != END_TSO_QUEUE;
3501 ASSERT(tso!=NULL && tso!=END_TSO_QUEUE); // sanity check
3502 ASSERT(get_itbl(tso)->type == TSO); // guess what, sanity check
3503 fprintf(stderr," TSO %d (%p),", tso->id, tso);
3505 fputc('\n', stderr);
3516 for (i=0, tso=run_queue_hd;
3517 tso != END_TSO_QUEUE;
3526 sched_belch(char *s, ...)
3530 #ifdef RTS_SUPPORTS_THREADS
3531 fprintf(stderr, "sched (task %p): ", osThreadId());
3533 fprintf(stderr, "== ");
3535 fprintf(stderr, "sched: ");
3537 vfprintf(stderr, s, ap);
3538 fprintf(stderr, "\n");