1 /* ---------------------------------------------------------------------------
2 * $Id: Schedule.c,v 1.147 2002/07/10 09:28:56 simonmar Exp $
4 * (c) The GHC Team, 1998-2000
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 distributed 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 * --------------------------------------------------------------------------*/
20 //@node Main scheduling code, , ,
21 //@section Main scheduling code
24 * Version with scheduler monitor support for SMPs (WAY=s):
26 This design provides a high-level API to create and schedule threads etc.
27 as documented in the SMP design document.
29 It uses a monitor design controlled by a single mutex to exercise control
30 over accesses to shared data structures, and builds on the Posix threads
33 The majority of state is shared. In order to keep essential per-task state,
34 there is a Capability structure, which contains all the information
35 needed to run a thread: its STG registers, a pointer to its TSO, a
36 nursery etc. During STG execution, a pointer to the capability is
37 kept in a register (BaseReg).
39 In a non-SMP build, there is one global capability, namely MainRegTable.
43 * Version with support for distributed memory parallelism aka GUM (WAY=mp):
45 The main scheduling loop in GUM iterates until a finish message is received.
46 In that case a global flag @receivedFinish@ is set and this instance of
47 the RTS shuts down. See ghc/rts/parallel/HLComms.c:processMessages()
48 for the handling of incoming messages, such as PP_FINISH.
49 Note that in the parallel case we have a system manager that coordinates
50 different PEs, each of which are running one instance of the RTS.
51 See ghc/rts/parallel/SysMan.c for the main routine of the parallel program.
52 From this routine processes executing ghc/rts/Main.c are spawned. -- HWL
54 * Version with support for simulating parallel execution aka GranSim (WAY=mg):
56 The main scheduling code in GranSim is quite different from that in std
57 (concurrent) Haskell: while concurrent Haskell just iterates over the
58 threads in the runnable queue, GranSim is event driven, i.e. it iterates
59 over the events in the global event queue. -- HWL
64 //* Variables and Data structures::
65 //* Main scheduling loop::
66 //* Suspend and Resume::
68 //* Garbage Collextion Routines::
69 //* Blocking Queue Routines::
70 //* Exception Handling Routines::
71 //* Debugging Routines::
75 //@node Includes, Variables and Data structures, Main scheduling code, Main scheduling code
76 //@subsection Includes
78 #include "PosixSource.h"
85 #include "StgStartup.h"
88 #include "StgMiscClosures.h"
90 #include "Interpreter.h"
91 #include "Exception.h"
99 #include "ThreadLabels.h"
101 #include "Proftimer.h"
102 #include "ProfHeap.h"
104 #if defined(GRAN) || defined(PAR)
105 # include "GranSimRts.h"
106 # include "GranSim.h"
107 # include "ParallelRts.h"
108 # include "Parallel.h"
109 # include "ParallelDebug.h"
110 # include "FetchMe.h"
114 #include "Capability.h"
115 #include "OSThreads.h"
118 #ifdef HAVE_SYS_TYPES_H
119 #include <sys/types.h>
127 //@node Variables and Data structures, Prototypes, Includes, Main scheduling code
128 //@subsection Variables and Data structures
130 /* Main thread queue.
131 * Locks required: sched_mutex.
133 StgMainThread *main_threads;
136 * Locks required: sched_mutex.
140 StgTSO* ActiveTSO = NULL; /* for assigning system costs; GranSim-Light only */
141 /* rtsTime TimeOfNextEvent, EndOfTimeSlice; now in GranSim.c */
144 In GranSim we have a runnable and a blocked queue for each processor.
145 In order to minimise code changes new arrays run_queue_hds/tls
146 are created. run_queue_hd is then a short cut (macro) for
147 run_queue_hds[CurrentProc] (see GranSim.h).
150 StgTSO *run_queue_hds[MAX_PROC], *run_queue_tls[MAX_PROC];
151 StgTSO *blocked_queue_hds[MAX_PROC], *blocked_queue_tls[MAX_PROC];
152 StgTSO *ccalling_threadss[MAX_PROC];
153 /* We use the same global list of threads (all_threads) in GranSim as in
154 the std RTS (i.e. we are cheating). However, we don't use this list in
155 the GranSim specific code at the moment (so we are only potentially
160 StgTSO *run_queue_hd, *run_queue_tl;
161 StgTSO *blocked_queue_hd, *blocked_queue_tl;
162 StgTSO *sleeping_queue; /* perhaps replace with a hash table? */
166 /* Linked list of all threads.
167 * Used for detecting garbage collected threads.
171 /* When a thread performs a safe C call (_ccall_GC, using old
172 * terminology), it gets put on the suspended_ccalling_threads
173 * list. Used by the garbage collector.
175 static StgTSO *suspended_ccalling_threads;
177 static StgTSO *threadStackOverflow(StgTSO *tso);
179 /* KH: The following two flags are shared memory locations. There is no need
180 to lock them, since they are only unset at the end of a scheduler
184 /* flag set by signal handler to precipitate a context switch */
185 //@cindex context_switch
188 /* if this flag is set as well, give up execution */
189 //@cindex interrupted
192 /* Next thread ID to allocate.
193 * Locks required: thread_id_mutex
195 //@cindex next_thread_id
196 StgThreadID next_thread_id = 1;
199 * Pointers to the state of the current thread.
200 * Rule of thumb: if CurrentTSO != NULL, then we're running a Haskell
201 * thread. If CurrentTSO == NULL, then we're at the scheduler level.
204 /* The smallest stack size that makes any sense is:
205 * RESERVED_STACK_WORDS (so we can get back from the stack overflow)
206 * + sizeofW(StgStopFrame) (the stg_stop_thread_info frame)
207 * + 1 (the realworld token for an IO thread)
208 * + 1 (the closure to enter)
210 * A thread with this stack will bomb immediately with a stack
211 * overflow, which will increase its stack size.
214 #define MIN_STACK_WORDS (RESERVED_STACK_WORDS + sizeofW(StgStopFrame) + 2)
221 /* This is used in `TSO.h' and gcc 2.96 insists that this variable actually
222 * exists - earlier gccs apparently didn't.
230 * Set to TRUE when entering a shutdown state (via shutdownHaskellAndExit()) --
231 * in an MT setting, needed to signal that a worker thread shouldn't hang around
232 * in the scheduler when it is out of work.
234 static rtsBool shutting_down_scheduler = rtsFalse;
236 void addToBlockedQueue ( StgTSO *tso );
238 static void schedule ( void );
239 void interruptStgRts ( void );
241 static void detectBlackHoles ( void );
244 static void sched_belch(char *s, ...);
247 #if defined(RTS_SUPPORTS_THREADS)
248 /* ToDo: carefully document the invariants that go together
249 * with these synchronisation objects.
251 Mutex sched_mutex = INIT_MUTEX_VAR;
252 Mutex term_mutex = INIT_MUTEX_VAR;
255 * A heavyweight solution to the problem of protecting
256 * the thread_id from concurrent update.
258 Mutex thread_id_mutex = INIT_MUTEX_VAR;
262 static Condition gc_pending_cond = INIT_COND_VAR;
266 #endif /* RTS_SUPPORTS_THREADS */
270 rtsTime TimeOfLastYield;
271 rtsBool emitSchedule = rtsTrue;
275 char *whatNext_strs[] = {
283 char *threadReturnCode_strs[] = {
284 "HeapOverflow", /* might also be StackOverflow */
293 StgTSO * createSparkThread(rtsSpark spark);
294 StgTSO * activateSpark (rtsSpark spark);
298 * The thread state for the main thread.
299 // ToDo: check whether not needed any more
303 #if defined(PAR) || defined(RTS_SUPPORTS_THREADS)
304 static void taskStart(void);
315 //@node Main scheduling loop, Suspend and Resume, Prototypes, Main scheduling code
316 //@subsection Main scheduling loop
318 /* ---------------------------------------------------------------------------
319 Main scheduling loop.
321 We use round-robin scheduling, each thread returning to the
322 scheduler loop when one of these conditions is detected:
325 * timer expires (thread yields)
330 Locking notes: we acquire the scheduler lock once at the beginning
331 of the scheduler loop, and release it when
333 * running a thread, or
334 * waiting for work, or
335 * waiting for a GC to complete.
338 In a GranSim setup this loop iterates over the global event queue.
339 This revolves around the global event queue, which determines what
340 to do next. Therefore, it's more complicated than either the
341 concurrent or the parallel (GUM) setup.
344 GUM iterates over incoming messages.
345 It starts with nothing to do (thus CurrentTSO == END_TSO_QUEUE),
346 and sends out a fish whenever it has nothing to do; in-between
347 doing the actual reductions (shared code below) it processes the
348 incoming messages and deals with delayed operations
349 (see PendingFetches).
350 This is not the ugliest code you could imagine, but it's bloody close.
352 ------------------------------------------------------------------------ */
359 StgThreadReturnCode ret;
367 rtsBool receivedFinish = rtsFalse;
369 nat tp_size, sp_size; // stats only
372 rtsBool was_interrupted = rtsFalse;
374 ACQUIRE_LOCK(&sched_mutex);
376 #if defined(RTS_SUPPORTS_THREADS)
377 waitForWorkCapability(&sched_mutex, &cap, rtsFalse);
379 /* simply initialise it in the non-threaded case */
380 grabCapability(&cap);
384 /* set up first event to get things going */
385 /* ToDo: assign costs for system setup and init MainTSO ! */
386 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
388 CurrentTSO, (StgClosure*)NULL, (rtsSpark*)NULL);
391 fprintf(stderr, "GRAN: Init CurrentTSO (in schedule) = %p\n", CurrentTSO);
392 G_TSO(CurrentTSO, 5));
394 if (RtsFlags.GranFlags.Light) {
395 /* Save current time; GranSim Light only */
396 CurrentTSO->gran.clock = CurrentTime[CurrentProc];
399 event = get_next_event();
401 while (event!=(rtsEvent*)NULL) {
402 /* Choose the processor with the next event */
403 CurrentProc = event->proc;
404 CurrentTSO = event->tso;
408 while (!receivedFinish) { /* set by processMessages */
409 /* when receiving PP_FINISH message */
416 IF_DEBUG(scheduler, printAllThreads());
418 #if defined(RTS_SUPPORTS_THREADS)
419 /* Check to see whether there are any worker threads
420 waiting to deposit external call results. If so,
421 yield our capability */
422 yieldToReturningWorker(&sched_mutex, &cap);
425 /* If we're interrupted (the user pressed ^C, or some other
426 * termination condition occurred), kill all the currently running
430 IF_DEBUG(scheduler, sched_belch("interrupted"));
432 interrupted = rtsFalse;
433 was_interrupted = rtsTrue;
436 /* Go through the list of main threads and wake up any
437 * clients whose computations have finished. ToDo: this
438 * should be done more efficiently without a linear scan
439 * of the main threads list, somehow...
441 #if defined(RTS_SUPPORTS_THREADS)
443 StgMainThread *m, **prev;
444 prev = &main_threads;
445 for (m = main_threads; m != NULL; m = m->link) {
446 switch (m->tso->what_next) {
449 *(m->ret) = (StgClosure *)m->tso->sp[0];
453 broadcastCondition(&m->wakeup);
455 removeThreadLabel(m->tso);
459 if (m->ret) *(m->ret) = NULL;
461 if (was_interrupted) {
462 m->stat = Interrupted;
466 broadcastCondition(&m->wakeup);
468 removeThreadLabel(m->tso);
477 #else /* not threaded */
480 /* in GUM do this only on the Main PE */
483 /* If our main thread has finished or been killed, return.
486 StgMainThread *m = main_threads;
487 if (m->tso->what_next == ThreadComplete
488 || m->tso->what_next == ThreadKilled) {
490 removeThreadLabel((StgWord)m->tso);
492 main_threads = main_threads->link;
493 if (m->tso->what_next == ThreadComplete) {
494 /* we finished successfully, fill in the return value */
495 if (m->ret) { *(m->ret) = (StgClosure *)m->tso->sp[0]; };
499 if (m->ret) { *(m->ret) = NULL; };
500 if (was_interrupted) {
501 m->stat = Interrupted;
511 /* Top up the run queue from our spark pool. We try to make the
512 * number of threads in the run queue equal to the number of
515 * Disable spark support in SMP for now, non-essential & requires
516 * a little bit of work to make it compile cleanly. -- sof 1/02.
518 #if 0 /* defined(SMP) */
520 nat n = getFreeCapabilities();
521 StgTSO *tso = run_queue_hd;
523 /* Count the run queue */
524 while (n > 0 && tso != END_TSO_QUEUE) {
531 spark = findSpark(rtsFalse);
533 break; /* no more sparks in the pool */
535 /* I'd prefer this to be done in activateSpark -- HWL */
536 /* tricky - it needs to hold the scheduler lock and
537 * not try to re-acquire it -- SDM */
538 createSparkThread(spark);
540 sched_belch("==^^ turning spark of closure %p into a thread",
541 (StgClosure *)spark));
544 /* We need to wake up the other tasks if we just created some
547 if (getFreeCapabilities() - n > 1) {
548 signalCondition( &thread_ready_cond );
553 /* check for signals each time around the scheduler */
554 #ifndef mingw32_TARGET_OS
555 if (signals_pending()) {
556 RELEASE_LOCK(&sched_mutex); /* ToDo: kill */
557 startSignalHandlers();
558 ACQUIRE_LOCK(&sched_mutex);
562 /* Check whether any waiting threads need to be woken up. If the
563 * run queue is empty, and there are no other tasks running, we
564 * can wait indefinitely for something to happen.
565 * ToDo: what if another client comes along & requests another
568 if ( !EMPTY_QUEUE(blocked_queue_hd) || !EMPTY_QUEUE(sleeping_queue) ) {
569 awaitEvent( EMPTY_RUN_QUEUE()
571 && allFreeCapabilities()
575 /* we can be interrupted while waiting for I/O... */
576 if (interrupted) continue;
579 * Detect deadlock: when we have no threads to run, there are no
580 * threads waiting on I/O or sleeping, and all the other tasks are
581 * waiting for work, we must have a deadlock of some description.
583 * We first try to find threads blocked on themselves (ie. black
584 * holes), and generate NonTermination exceptions where necessary.
586 * If no threads are black holed, we have a deadlock situation, so
587 * inform all the main threads.
590 if ( EMPTY_THREAD_QUEUES()
591 #if defined(RTS_SUPPORTS_THREADS)
592 && EMPTY_QUEUE(suspended_ccalling_threads)
595 && allFreeCapabilities()
599 IF_DEBUG(scheduler, sched_belch("deadlocked, forcing major GC..."));
600 #if defined(THREADED_RTS)
601 /* and SMP mode ..? */
602 releaseCapability(cap);
604 // Garbage collection can release some new threads due to
605 // either (a) finalizers or (b) threads resurrected because
606 // they are about to be send BlockedOnDeadMVar. Any threads
607 // thus released will be immediately runnable.
608 GarbageCollect(GetRoots,rtsTrue);
610 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
613 sched_belch("still deadlocked, checking for black holes..."));
616 if ( !EMPTY_RUN_QUEUE() ) { goto not_deadlocked; }
618 #ifndef mingw32_TARGET_OS
619 /* If we have user-installed signal handlers, then wait
620 * for signals to arrive rather then bombing out with a
623 #if defined(RTS_SUPPORTS_THREADS)
624 if ( 0 ) { /* hmm..what to do? Simply stop waiting for
625 a signal with no runnable threads (or I/O
626 suspended ones) leads nowhere quick.
627 For now, simply shut down when we reach this
630 ToDo: define precisely under what conditions
631 the Scheduler should shut down in an MT setting.
634 if ( anyUserHandlers() ) {
637 sched_belch("still deadlocked, waiting for signals..."));
641 // we might be interrupted...
642 if (interrupted) { continue; }
644 if (signals_pending()) {
645 RELEASE_LOCK(&sched_mutex);
646 startSignalHandlers();
647 ACQUIRE_LOCK(&sched_mutex);
649 ASSERT(!EMPTY_RUN_QUEUE());
654 /* Probably a real deadlock. Send the current main thread the
655 * Deadlock exception (or in the SMP build, send *all* main
656 * threads the deadlock exception, since none of them can make
661 #if defined(RTS_SUPPORTS_THREADS)
662 for (m = main_threads; m != NULL; m = m->link) {
663 switch (m->tso->why_blocked) {
664 case BlockedOnBlackHole:
665 raiseAsync(m->tso, (StgClosure *)NonTermination_closure);
667 case BlockedOnException:
669 raiseAsync(m->tso, (StgClosure *)Deadlock_closure);
672 barf("deadlock: main thread blocked in a strange way");
677 switch (m->tso->why_blocked) {
678 case BlockedOnBlackHole:
679 raiseAsync(m->tso, (StgClosure *)NonTermination_closure);
681 case BlockedOnException:
683 raiseAsync(m->tso, (StgClosure *)Deadlock_closure);
686 barf("deadlock: main thread blocked in a strange way");
691 #if defined(RTS_SUPPORTS_THREADS)
692 /* ToDo: revisit conditions (and mechanism) for shutting
693 down a multi-threaded world */
694 IF_DEBUG(scheduler, sched_belch("all done, i think...shutting down."));
695 RELEASE_LOCK(&sched_mutex);
703 /* ToDo: add deadlock detection in GUM (similar to SMP) -- HWL */
707 /* If there's a GC pending, don't do anything until it has
711 IF_DEBUG(scheduler,sched_belch("waiting for GC"));
712 waitCondition( &gc_pending_cond, &sched_mutex );
716 #if defined(RTS_SUPPORTS_THREADS)
717 /* block until we've got a thread on the run queue and a free
721 if ( EMPTY_RUN_QUEUE() ) {
722 /* Give up our capability */
723 releaseCapability(cap);
725 /* If we're in the process of shutting down (& running the
726 * a batch of finalisers), don't wait around.
728 if ( shutting_down_scheduler ) {
729 RELEASE_LOCK(&sched_mutex);
732 IF_DEBUG(scheduler, sched_belch("thread %d: waiting for work", osThreadId()));
733 waitForWorkCapability(&sched_mutex, &cap, rtsTrue);
734 IF_DEBUG(scheduler, sched_belch("thread %d: work now available", osThreadId()));
739 if (RtsFlags.GranFlags.Light)
740 GranSimLight_enter_system(event, &ActiveTSO); // adjust ActiveTSO etc
742 /* adjust time based on time-stamp */
743 if (event->time > CurrentTime[CurrentProc] &&
744 event->evttype != ContinueThread)
745 CurrentTime[CurrentProc] = event->time;
747 /* Deal with the idle PEs (may issue FindWork or MoveSpark events) */
748 if (!RtsFlags.GranFlags.Light)
751 IF_DEBUG(gran, fprintf(stderr, "GRAN: switch by event-type\n"));
753 /* main event dispatcher in GranSim */
754 switch (event->evttype) {
755 /* Should just be continuing execution */
757 IF_DEBUG(gran, fprintf(stderr, "GRAN: doing ContinueThread\n"));
758 /* ToDo: check assertion
759 ASSERT(run_queue_hd != (StgTSO*)NULL &&
760 run_queue_hd != END_TSO_QUEUE);
762 /* Ignore ContinueThreads for fetching threads (if synchr comm) */
763 if (!RtsFlags.GranFlags.DoAsyncFetch &&
764 procStatus[CurrentProc]==Fetching) {
765 belch("ghuH: Spurious ContinueThread while Fetching ignored; TSO %d (%p) [PE %d]",
766 CurrentTSO->id, CurrentTSO, CurrentProc);
769 /* Ignore ContinueThreads for completed threads */
770 if (CurrentTSO->what_next == ThreadComplete) {
771 belch("ghuH: found a ContinueThread event for completed thread %d (%p) [PE %d] (ignoring ContinueThread)",
772 CurrentTSO->id, CurrentTSO, CurrentProc);
775 /* Ignore ContinueThreads for threads that are being migrated */
776 if (PROCS(CurrentTSO)==Nowhere) {
777 belch("ghuH: trying to run the migrating TSO %d (%p) [PE %d] (ignoring ContinueThread)",
778 CurrentTSO->id, CurrentTSO, CurrentProc);
781 /* The thread should be at the beginning of the run queue */
782 if (CurrentTSO!=run_queue_hds[CurrentProc]) {
783 belch("ghuH: TSO %d (%p) [PE %d] is not at the start of the run_queue when doing a ContinueThread",
784 CurrentTSO->id, CurrentTSO, CurrentProc);
785 break; // run the thread anyway
788 new_event(proc, proc, CurrentTime[proc],
790 (StgTSO*)NULL, (StgClosure*)NULL, (rtsSpark*)NULL);
792 */ /* Catches superfluous CONTINUEs -- should be unnecessary */
793 break; // now actually run the thread; DaH Qu'vam yImuHbej
796 do_the_fetchnode(event);
797 goto next_thread; /* handle next event in event queue */
800 do_the_globalblock(event);
801 goto next_thread; /* handle next event in event queue */
804 do_the_fetchreply(event);
805 goto next_thread; /* handle next event in event queue */
807 case UnblockThread: /* Move from the blocked queue to the tail of */
808 do_the_unblock(event);
809 goto next_thread; /* handle next event in event queue */
811 case ResumeThread: /* Move from the blocked queue to the tail of */
812 /* the runnable queue ( i.e. Qu' SImqa'lu') */
813 event->tso->gran.blocktime +=
814 CurrentTime[CurrentProc] - event->tso->gran.blockedat;
815 do_the_startthread(event);
816 goto next_thread; /* handle next event in event queue */
819 do_the_startthread(event);
820 goto next_thread; /* handle next event in event queue */
823 do_the_movethread(event);
824 goto next_thread; /* handle next event in event queue */
827 do_the_movespark(event);
828 goto next_thread; /* handle next event in event queue */
831 do_the_findwork(event);
832 goto next_thread; /* handle next event in event queue */
835 barf("Illegal event type %u\n", event->evttype);
838 /* This point was scheduler_loop in the old RTS */
840 IF_DEBUG(gran, belch("GRAN: after main switch"));
842 TimeOfLastEvent = CurrentTime[CurrentProc];
843 TimeOfNextEvent = get_time_of_next_event();
844 IgnoreEvents=(TimeOfNextEvent==0); // HWL HACK
845 // CurrentTSO = ThreadQueueHd;
847 IF_DEBUG(gran, belch("GRAN: time of next event is: %ld",
850 if (RtsFlags.GranFlags.Light)
851 GranSimLight_leave_system(event, &ActiveTSO);
853 EndOfTimeSlice = CurrentTime[CurrentProc]+RtsFlags.GranFlags.time_slice;
856 belch("GRAN: end of time-slice is %#lx", EndOfTimeSlice));
858 /* in a GranSim setup the TSO stays on the run queue */
860 /* Take a thread from the run queue. */
861 t = POP_RUN_QUEUE(); // take_off_run_queue(t);
864 fprintf(stderr, "GRAN: About to run current thread, which is\n");
867 context_switch = 0; // turned on via GranYield, checking events and time slice
870 DumpGranEvent(GR_SCHEDULE, t));
872 procStatus[CurrentProc] = Busy;
875 if (PendingFetches != END_BF_QUEUE) {
879 /* ToDo: phps merge with spark activation above */
880 /* check whether we have local work and send requests if we have none */
881 if (EMPTY_RUN_QUEUE()) { /* no runnable threads */
882 /* :-[ no local threads => look out for local sparks */
883 /* the spark pool for the current PE */
884 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
885 if (advisory_thread_count < RtsFlags.ParFlags.maxThreads &&
886 pool->hd < pool->tl) {
888 * ToDo: add GC code check that we really have enough heap afterwards!!
890 * If we're here (no runnable threads) and we have pending
891 * sparks, we must have a space problem. Get enough space
892 * to turn one of those pending sparks into a
896 spark = findSpark(rtsFalse); /* get a spark */
897 if (spark != (rtsSpark) NULL) {
898 tso = activateSpark(spark); /* turn the spark into a thread */
899 IF_PAR_DEBUG(schedule,
900 belch("==== schedule: Created TSO %d (%p); %d threads active",
901 tso->id, tso, advisory_thread_count));
903 if (tso==END_TSO_QUEUE) { /* failed to activate spark->back to loop */
904 belch("==^^ failed to activate spark");
906 } /* otherwise fall through & pick-up new tso */
908 IF_PAR_DEBUG(verbose,
909 belch("==^^ no local sparks (spark pool contains only NFs: %d)",
910 spark_queue_len(pool)));
915 /* If we still have no work we need to send a FISH to get a spark
918 if (EMPTY_RUN_QUEUE()) {
919 /* =8-[ no local sparks => look for work on other PEs */
921 * We really have absolutely no work. Send out a fish
922 * (there may be some out there already), and wait for
923 * something to arrive. We clearly can't run any threads
924 * until a SCHEDULE or RESUME arrives, and so that's what
925 * we're hoping to see. (Of course, we still have to
926 * respond to other types of messages.)
928 TIME now = msTime() /*CURRENT_TIME*/;
929 IF_PAR_DEBUG(verbose,
930 belch("-- now=%ld", now));
931 IF_PAR_DEBUG(verbose,
932 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
933 (last_fish_arrived_at!=0 &&
934 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay > now)) {
935 belch("--$$ delaying FISH until %ld (last fish %ld, delay %ld, now %ld)",
936 last_fish_arrived_at+RtsFlags.ParFlags.fishDelay,
937 last_fish_arrived_at,
938 RtsFlags.ParFlags.fishDelay, now);
941 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
942 (last_fish_arrived_at==0 ||
943 (last_fish_arrived_at+RtsFlags.ParFlags.fishDelay <= now))) {
944 /* outstandingFishes is set in sendFish, processFish;
945 avoid flooding system with fishes via delay */
947 sendFish(pe, mytid, NEW_FISH_AGE, NEW_FISH_HISTORY,
950 // Global statistics: count no. of fishes
951 if (RtsFlags.ParFlags.ParStats.Global &&
952 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
953 globalParStats.tot_fish_mess++;
957 receivedFinish = processMessages();
960 } else if (PacketsWaiting()) { /* Look for incoming messages */
961 receivedFinish = processMessages();
964 /* Now we are sure that we have some work available */
965 ASSERT(run_queue_hd != END_TSO_QUEUE);
967 /* Take a thread from the run queue, if we have work */
968 t = POP_RUN_QUEUE(); // take_off_run_queue(END_TSO_QUEUE);
969 IF_DEBUG(sanity,checkTSO(t));
971 /* ToDo: write something to the log-file
972 if (RTSflags.ParFlags.granSimStats && !sameThread)
973 DumpGranEvent(GR_SCHEDULE, RunnableThreadsHd);
977 /* the spark pool for the current PE */
978 pool = &(MainRegTable.rSparks); // generalise to cap = &MainRegTable
981 belch("--=^ %d threads, %d sparks on [%#x]",
982 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
985 if (0 && RtsFlags.ParFlags.ParStats.Full &&
986 t && LastTSO && t->id != LastTSO->id &&
987 LastTSO->why_blocked == NotBlocked &&
988 LastTSO->what_next != ThreadComplete) {
989 // if previously scheduled TSO not blocked we have to record the context switch
990 DumpVeryRawGranEvent(TimeOfLastYield, CURRENT_PROC, CURRENT_PROC,
991 GR_DESCHEDULE, LastTSO, (StgClosure *)NULL, 0, 0);
994 if (RtsFlags.ParFlags.ParStats.Full &&
995 (emitSchedule /* forced emit */ ||
996 (t && LastTSO && t->id != LastTSO->id))) {
998 we are running a different TSO, so write a schedule event to log file
999 NB: If we use fair scheduling we also have to write a deschedule
1000 event for LastTSO; with unfair scheduling we know that the
1001 previous tso has blocked whenever we switch to another tso, so
1002 we don't need it in GUM for now
1004 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1005 GR_SCHEDULE, t, (StgClosure *)NULL, 0, 0);
1006 emitSchedule = rtsFalse;
1010 #else /* !GRAN && !PAR */
1012 /* grab a thread from the run queue */
1013 ASSERT(run_queue_hd != END_TSO_QUEUE);
1014 t = POP_RUN_QUEUE();
1015 // Sanity check the thread we're about to run. This can be
1016 // expensive if there is lots of thread switching going on...
1017 IF_DEBUG(sanity,checkTSO(t));
1020 cap->r.rCurrentTSO = t;
1022 /* context switches are now initiated by the timer signal, unless
1023 * the user specified "context switch as often as possible", with
1028 RtsFlags.ProfFlags.profileInterval == 0 ||
1030 (RtsFlags.ConcFlags.ctxtSwitchTicks == 0
1031 && (run_queue_hd != END_TSO_QUEUE
1032 || blocked_queue_hd != END_TSO_QUEUE
1033 || sleeping_queue != END_TSO_QUEUE)))
1038 RELEASE_LOCK(&sched_mutex);
1040 IF_DEBUG(scheduler, sched_belch("-->> Running TSO %ld (%p) %s ...",
1041 t->id, t, whatNext_strs[t->what_next]));
1044 startHeapProfTimer();
1047 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
1048 /* Run the current thread
1050 switch (cap->r.rCurrentTSO->what_next) {
1052 case ThreadComplete:
1053 /* Thread already finished, return to scheduler. */
1054 ret = ThreadFinished;
1056 case ThreadEnterGHC:
1057 ret = StgRun((StgFunPtr) stg_enterStackTop, &cap->r);
1060 ret = StgRun((StgFunPtr) stg_returnToStackTop, &cap->r);
1062 case ThreadEnterInterp:
1063 ret = interpretBCO(cap);
1066 barf("schedule: invalid what_next field");
1068 /* +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ */
1070 /* Costs for the scheduler are assigned to CCS_SYSTEM */
1072 stopHeapProfTimer();
1076 ACQUIRE_LOCK(&sched_mutex);
1079 IF_DEBUG(scheduler,fprintf(stderr,"scheduler (task %ld): ", osThreadId()););
1080 #elif !defined(GRAN) && !defined(PAR)
1081 IF_DEBUG(scheduler,fprintf(stderr,"scheduler: "););
1083 t = cap->r.rCurrentTSO;
1086 /* HACK 675: if the last thread didn't yield, make sure to print a
1087 SCHEDULE event to the log file when StgRunning the next thread, even
1088 if it is the same one as before */
1090 TimeOfLastYield = CURRENT_TIME;
1096 IF_DEBUG(gran, DumpGranEvent(GR_DESCHEDULE, t));
1097 globalGranStats.tot_heapover++;
1099 globalParStats.tot_heapover++;
1102 // did the task ask for a large block?
1103 if (cap->r.rHpAlloc > BLOCK_SIZE_W) {
1104 // if so, get one and push it on the front of the nursery.
1108 blocks = (nat)BLOCK_ROUND_UP(cap->r.rHpAlloc * sizeof(W_)) / BLOCK_SIZE;
1110 IF_DEBUG(scheduler,belch("--<< thread %ld (%p; %s) stopped: requesting a large block (size %d)",
1112 whatNext_strs[t->what_next], blocks));
1114 // don't do this if it would push us over the
1115 // alloc_blocks_lim limit; we'll GC first.
1116 if (alloc_blocks + blocks < alloc_blocks_lim) {
1118 alloc_blocks += blocks;
1119 bd = allocGroup( blocks );
1121 // link the new group into the list
1122 bd->link = cap->r.rCurrentNursery;
1123 bd->u.back = cap->r.rCurrentNursery->u.back;
1124 if (cap->r.rCurrentNursery->u.back != NULL) {
1125 cap->r.rCurrentNursery->u.back->link = bd;
1127 ASSERT(g0s0->blocks == cap->r.rCurrentNursery &&
1128 g0s0->blocks == cap->r.rNursery);
1129 cap->r.rNursery = g0s0->blocks = bd;
1131 cap->r.rCurrentNursery->u.back = bd;
1133 // initialise it as a nursery block. We initialise the
1134 // step, gen_no, and flags field of *every* sub-block in
1135 // this large block, because this is easier than making
1136 // sure that we always find the block head of a large
1137 // block whenever we call Bdescr() (eg. evacuate() and
1138 // isAlive() in the GC would both have to do this, at
1142 for (x = bd; x < bd + blocks; x++) {
1150 // don't forget to update the block count in g0s0.
1151 g0s0->n_blocks += blocks;
1152 ASSERT(countBlocks(g0s0->blocks) == g0s0->n_blocks);
1154 // now update the nursery to point to the new block
1155 cap->r.rCurrentNursery = bd;
1157 // we might be unlucky and have another thread get on the
1158 // run queue before us and steal the large block, but in that
1159 // case the thread will just end up requesting another large
1161 PUSH_ON_RUN_QUEUE(t);
1166 /* make all the running tasks block on a condition variable,
1167 * maybe set context_switch and wait till they all pile in,
1168 * then have them wait on a GC condition variable.
1170 IF_DEBUG(scheduler,belch("--<< thread %ld (%p; %s) stopped: HeapOverflow",
1171 t->id, t, whatNext_strs[t->what_next]));
1174 ASSERT(!is_on_queue(t,CurrentProc));
1176 /* Currently we emit a DESCHEDULE event before GC in GUM.
1177 ToDo: either add separate event to distinguish SYSTEM time from rest
1178 or just nuke this DESCHEDULE (and the following SCHEDULE) */
1179 if (0 && RtsFlags.ParFlags.ParStats.Full) {
1180 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1181 GR_DESCHEDULE, t, (StgClosure *)NULL, 0, 0);
1182 emitSchedule = rtsTrue;
1186 ready_to_gc = rtsTrue;
1187 context_switch = 1; /* stop other threads ASAP */
1188 PUSH_ON_RUN_QUEUE(t);
1189 /* actual GC is done at the end of the while loop */
1195 DumpGranEvent(GR_DESCHEDULE, t));
1196 globalGranStats.tot_stackover++;
1199 // DumpGranEvent(GR_DESCHEDULE, t);
1200 globalParStats.tot_stackover++;
1202 IF_DEBUG(scheduler,belch("--<< thread %ld (%p; %s) stopped, StackOverflow",
1203 t->id, t, whatNext_strs[t->what_next]));
1204 /* just adjust the stack for this thread, then pop it back
1210 /* enlarge the stack */
1211 StgTSO *new_t = threadStackOverflow(t);
1213 /* This TSO has moved, so update any pointers to it from the
1214 * main thread stack. It better not be on any other queues...
1215 * (it shouldn't be).
1217 for (m = main_threads; m != NULL; m = m->link) {
1222 threadPaused(new_t);
1223 PUSH_ON_RUN_QUEUE(new_t);
1227 case ThreadYielding:
1230 DumpGranEvent(GR_DESCHEDULE, t));
1231 globalGranStats.tot_yields++;
1234 // DumpGranEvent(GR_DESCHEDULE, t);
1235 globalParStats.tot_yields++;
1237 /* put the thread back on the run queue. Then, if we're ready to
1238 * GC, check whether this is the last task to stop. If so, wake
1239 * up the GC thread. getThread will block during a GC until the
1243 if (t->what_next == ThreadEnterInterp) {
1244 /* ToDo: or maybe a timer expired when we were in Hugs?
1245 * or maybe someone hit ctrl-C
1247 belch("--<< thread %ld (%p; %s) stopped to switch to Hugs",
1248 t->id, t, whatNext_strs[t->what_next]);
1250 belch("--<< thread %ld (%p; %s) stopped, yielding",
1251 t->id, t, whatNext_strs[t->what_next]);
1258 //belch("&& Doing sanity check on yielding TSO %ld.", t->id);
1260 ASSERT(t->link == END_TSO_QUEUE);
1262 ASSERT(!is_on_queue(t,CurrentProc));
1265 //belch("&& Doing sanity check on all ThreadQueues (and their TSOs).");
1266 checkThreadQsSanity(rtsTrue));
1269 if (RtsFlags.ParFlags.doFairScheduling) {
1270 /* this does round-robin scheduling; good for concurrency */
1271 APPEND_TO_RUN_QUEUE(t);
1273 /* this does unfair scheduling; good for parallelism */
1274 PUSH_ON_RUN_QUEUE(t);
1277 /* this does round-robin scheduling; good for concurrency */
1278 APPEND_TO_RUN_QUEUE(t);
1281 /* add a ContinueThread event to actually process the thread */
1282 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1284 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1286 belch("GRAN: eventq and runnableq after adding yielded thread to queue again:");
1295 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: ",
1296 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)));
1297 if (t->block_info.closure!=(StgClosure*)NULL) print_bq(t->block_info.closure));
1299 // ??? needed; should emit block before
1301 DumpGranEvent(GR_DESCHEDULE, t));
1302 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1305 ASSERT(procStatus[CurrentProc]==Busy ||
1306 ((procStatus[CurrentProc]==Fetching) &&
1307 (t->block_info.closure!=(StgClosure*)NULL)));
1308 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1309 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1310 procStatus[CurrentProc]==Fetching))
1311 procStatus[CurrentProc] = Idle;
1315 belch("--<< thread %ld (%p; %s) stopped, blocking on node %p with BQ: ",
1316 t->id, t, whatNext_strs[t->what_next], t->block_info.closure));
1319 if (t->block_info.closure!=(StgClosure*)NULL)
1320 print_bq(t->block_info.closure));
1322 /* Send a fetch (if BlockedOnGA) and dump event to log file */
1325 /* whatever we schedule next, we must log that schedule */
1326 emitSchedule = rtsTrue;
1329 /* don't need to do anything. Either the thread is blocked on
1330 * I/O, in which case we'll have called addToBlockedQueue
1331 * previously, or it's blocked on an MVar or Blackhole, in which
1332 * case it'll be on the relevant queue already.
1335 fprintf(stderr, "--<< thread %d (%p) stopped: ", t->id, t);
1336 printThreadBlockage(t);
1337 fprintf(stderr, "\n"));
1339 /* Only for dumping event to log file
1340 ToDo: do I need this in GranSim, too?
1347 case ThreadFinished:
1348 /* Need to check whether this was a main thread, and if so, signal
1349 * the task that started it with the return value. If we have no
1350 * more main threads, we probably need to stop all the tasks until
1353 /* We also end up here if the thread kills itself with an
1354 * uncaught exception, see Exception.hc.
1356 IF_DEBUG(scheduler,belch("--++ thread %d (%p) finished", t->id, t));
1358 endThread(t, CurrentProc); // clean-up the thread
1360 /* For now all are advisory -- HWL */
1361 //if(t->priority==AdvisoryPriority) ??
1362 advisory_thread_count--;
1365 if(t->dist.priority==RevalPriority)
1369 if (RtsFlags.ParFlags.ParStats.Full &&
1370 !RtsFlags.ParFlags.ParStats.Suppressed)
1371 DumpEndEvent(CURRENT_PROC, t, rtsFalse /* not mandatory */);
1376 barf("schedule: invalid thread return code %d", (int)ret);
1380 if (RtsFlags.ProfFlags.profileInterval==0 || performHeapProfile) {
1381 GarbageCollect(GetRoots, rtsTrue);
1383 performHeapProfile = rtsFalse;
1384 ready_to_gc = rtsFalse; // we already GC'd
1390 && allFreeCapabilities()
1393 /* everybody back, start the GC.
1394 * Could do it in this thread, or signal a condition var
1395 * to do it in another thread. Either way, we need to
1396 * broadcast on gc_pending_cond afterward.
1398 #if defined(RTS_SUPPORTS_THREADS)
1399 IF_DEBUG(scheduler,sched_belch("doing GC"));
1401 GarbageCollect(GetRoots,rtsFalse);
1402 ready_to_gc = rtsFalse;
1404 broadcastCondition(&gc_pending_cond);
1407 /* add a ContinueThread event to continue execution of current thread */
1408 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1410 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1412 fprintf(stderr, "GRAN: eventq and runnableq after Garbage collection:\n");
1420 IF_GRAN_DEBUG(unused,
1421 print_eventq(EventHd));
1423 event = get_next_event();
1426 /* ToDo: wait for next message to arrive rather than busy wait */
1429 } /* end of while(1) */
1431 IF_PAR_DEBUG(verbose,
1432 belch("== Leaving schedule() after having received Finish"));
1435 /* ---------------------------------------------------------------------------
1436 * Singleton fork(). Do not copy any running threads.
1437 * ------------------------------------------------------------------------- */
1439 StgInt forkProcess(StgTSO* tso) {
1441 #ifndef mingw32_TARGET_OS
1445 IF_DEBUG(scheduler,sched_belch("forking!"));
1448 if (pid) { /* parent */
1450 /* just return the pid */
1452 } else { /* child */
1453 /* wipe all other threads */
1455 tso->link = END_TSO_QUEUE;
1457 /* DO NOT TOUCH THE QUEUES directly because most of the code around
1458 us is picky about finding the threat still in its queue when
1459 handling the deleteThread() */
1461 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1463 if (t->id != tso->id) {
1470 barf("forkProcess#: primop not implemented for mingw32, sorry! (%u)\n", tso->id);
1471 /* pointlessly printing out the TSOs 'id' to avoid CC unused warning. */
1473 #endif /* mingw32 */
1476 /* ---------------------------------------------------------------------------
1477 * deleteAllThreads(): kill all the live threads.
1479 * This is used when we catch a user interrupt (^C), before performing
1480 * any necessary cleanups and running finalizers.
1482 * Locks: sched_mutex held.
1483 * ------------------------------------------------------------------------- */
1485 void deleteAllThreads ( void )
1488 IF_DEBUG(scheduler,sched_belch("deleting all threads"));
1489 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
1490 next = t->global_link;
1493 run_queue_hd = run_queue_tl = END_TSO_QUEUE;
1494 blocked_queue_hd = blocked_queue_tl = END_TSO_QUEUE;
1495 sleeping_queue = END_TSO_QUEUE;
1498 /* startThread and insertThread are now in GranSim.c -- HWL */
1501 //@node Suspend and Resume, Run queue code, Main scheduling loop, Main scheduling code
1502 //@subsection Suspend and Resume
1504 /* ---------------------------------------------------------------------------
1505 * Suspending & resuming Haskell threads.
1507 * When making a "safe" call to C (aka _ccall_GC), the task gives back
1508 * its capability before calling the C function. This allows another
1509 * task to pick up the capability and carry on running Haskell
1510 * threads. It also means that if the C call blocks, it won't lock
1513 * The Haskell thread making the C call is put to sleep for the
1514 * duration of the call, on the susepended_ccalling_threads queue. We
1515 * give out a token to the task, which it can use to resume the thread
1516 * on return from the C function.
1517 * ------------------------------------------------------------------------- */
1520 suspendThread( StgRegTable *reg,
1522 #if !defined(RTS_SUPPORTS_THREADS) && !defined(DEBUG)
1530 /* assume that *reg is a pointer to the StgRegTable part
1533 cap = (Capability *)((void *)reg - sizeof(StgFunTable));
1535 ACQUIRE_LOCK(&sched_mutex);
1538 sched_belch("thread %d did a _ccall_gc (is_concurrent: %d)", cap->r.rCurrentTSO->id,concCall));
1540 threadPaused(cap->r.rCurrentTSO);
1541 cap->r.rCurrentTSO->link = suspended_ccalling_threads;
1542 suspended_ccalling_threads = cap->r.rCurrentTSO;
1544 #if defined(RTS_SUPPORTS_THREADS)
1545 cap->r.rCurrentTSO->why_blocked = BlockedOnCCall;
1548 /* Use the thread ID as the token; it should be unique */
1549 tok = cap->r.rCurrentTSO->id;
1551 /* Hand back capability */
1552 releaseCapability(cap);
1554 #if defined(RTS_SUPPORTS_THREADS)
1555 /* Preparing to leave the RTS, so ensure there's a native thread/task
1556 waiting to take over.
1558 ToDo: optimise this and only create a new task if there's a need
1559 for one (i.e., if there's only one Concurrent Haskell thread alive,
1560 there's no need to create a new task).
1562 IF_DEBUG(scheduler, sched_belch("worker thread (%d): leaving RTS", tok));
1564 startTask(taskStart);
1568 /* Other threads _might_ be available for execution; signal this */
1570 RELEASE_LOCK(&sched_mutex);
1575 resumeThread( StgInt tok,
1577 #if !defined(RTS_SUPPORTS_THREADS)
1582 StgTSO *tso, **prev;
1585 #if defined(RTS_SUPPORTS_THREADS)
1586 /* Wait for permission to re-enter the RTS with the result. */
1588 ACQUIRE_LOCK(&sched_mutex);
1589 grabReturnCapability(&sched_mutex, &cap);
1591 grabCapability(&cap);
1594 grabCapability(&cap);
1597 /* Remove the thread off of the suspended list */
1598 prev = &suspended_ccalling_threads;
1599 for (tso = suspended_ccalling_threads;
1600 tso != END_TSO_QUEUE;
1601 prev = &tso->link, tso = tso->link) {
1602 if (tso->id == (StgThreadID)tok) {
1607 if (tso == END_TSO_QUEUE) {
1608 barf("resumeThread: thread not found");
1610 tso->link = END_TSO_QUEUE;
1611 /* Reset blocking status */
1612 tso->why_blocked = NotBlocked;
1614 cap->r.rCurrentTSO = tso;
1615 RELEASE_LOCK(&sched_mutex);
1620 /* ---------------------------------------------------------------------------
1622 * ------------------------------------------------------------------------ */
1623 static void unblockThread(StgTSO *tso);
1625 /* ---------------------------------------------------------------------------
1626 * Comparing Thread ids.
1628 * This is used from STG land in the implementation of the
1629 * instances of Eq/Ord for ThreadIds.
1630 * ------------------------------------------------------------------------ */
1632 int cmp_thread(const StgTSO *tso1, const StgTSO *tso2)
1634 StgThreadID id1 = tso1->id;
1635 StgThreadID id2 = tso2->id;
1637 if (id1 < id2) return (-1);
1638 if (id1 > id2) return 1;
1642 /* ---------------------------------------------------------------------------
1643 * Fetching the ThreadID from an StgTSO.
1645 * This is used in the implementation of Show for ThreadIds.
1646 * ------------------------------------------------------------------------ */
1647 int rts_getThreadId(const StgTSO *tso)
1653 void labelThread(StgTSO *tso, char *label)
1658 /* Caveat: Once set, you can only set the thread name to "" */
1659 len = strlen(label)+1;
1662 fprintf(stderr,"insufficient memory for labelThread!\n");
1664 strncpy(buf,label,len);
1665 /* Update will free the old memory for us */
1666 updateThreadLabel((StgWord)tso,buf);
1670 /* ---------------------------------------------------------------------------
1671 Create a new thread.
1673 The new thread starts with the given stack size. Before the
1674 scheduler can run, however, this thread needs to have a closure
1675 (and possibly some arguments) pushed on its stack. See
1676 pushClosure() in Schedule.h.
1678 createGenThread() and createIOThread() (in SchedAPI.h) are
1679 convenient packaged versions of this function.
1681 currently pri (priority) is only used in a GRAN setup -- HWL
1682 ------------------------------------------------------------------------ */
1683 //@cindex createThread
1685 /* currently pri (priority) is only used in a GRAN setup -- HWL */
1687 createThread(nat size, StgInt pri)
1690 createThread(nat size)
1697 /* First check whether we should create a thread at all */
1699 /* check that no more than RtsFlags.ParFlags.maxThreads threads are created */
1700 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads) {
1702 belch("{createThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1703 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1704 return END_TSO_QUEUE;
1710 ASSERT(!RtsFlags.GranFlags.Light || CurrentProc==0);
1713 // ToDo: check whether size = stack_size - TSO_STRUCT_SIZEW
1715 /* catch ridiculously small stack sizes */
1716 if (size < MIN_STACK_WORDS + TSO_STRUCT_SIZEW) {
1717 size = MIN_STACK_WORDS + TSO_STRUCT_SIZEW;
1720 stack_size = size - TSO_STRUCT_SIZEW;
1722 tso = (StgTSO *)allocate(size);
1723 TICK_ALLOC_TSO(stack_size, 0);
1725 SET_HDR(tso, &stg_TSO_info, CCS_SYSTEM);
1727 SET_GRAN_HDR(tso, ThisPE);
1729 tso->what_next = ThreadEnterGHC;
1731 /* tso->id needs to be unique. For now we use a heavyweight mutex to
1732 * protect the increment operation on next_thread_id.
1733 * In future, we could use an atomic increment instead.
1735 ACQUIRE_LOCK(&thread_id_mutex);
1736 tso->id = next_thread_id++;
1737 RELEASE_LOCK(&thread_id_mutex);
1739 tso->why_blocked = NotBlocked;
1740 tso->blocked_exceptions = NULL;
1742 tso->stack_size = stack_size;
1743 tso->max_stack_size = round_to_mblocks(RtsFlags.GcFlags.maxStkSize)
1745 tso->sp = (P_)&(tso->stack) + stack_size;
1748 tso->prof.CCCS = CCS_MAIN;
1751 /* put a stop frame on the stack */
1752 tso->sp -= sizeofW(StgStopFrame);
1753 SET_HDR((StgClosure*)tso->sp,(StgInfoTable *)&stg_stop_thread_info,CCS_SYSTEM);
1754 tso->su = (StgUpdateFrame*)tso->sp;
1758 tso->link = END_TSO_QUEUE;
1759 /* uses more flexible routine in GranSim */
1760 insertThread(tso, CurrentProc);
1762 /* In a non-GranSim setup the pushing of a TSO onto the runq is separated
1768 if (RtsFlags.GranFlags.GranSimStats.Full)
1769 DumpGranEvent(GR_START,tso);
1771 if (RtsFlags.ParFlags.ParStats.Full)
1772 DumpGranEvent(GR_STARTQ,tso);
1773 /* HACk to avoid SCHEDULE
1777 /* Link the new thread on the global thread list.
1779 tso->global_link = all_threads;
1783 tso->dist.priority = MandatoryPriority; //by default that is...
1787 tso->gran.pri = pri;
1789 tso->gran.magic = TSO_MAGIC; // debugging only
1791 tso->gran.sparkname = 0;
1792 tso->gran.startedat = CURRENT_TIME;
1793 tso->gran.exported = 0;
1794 tso->gran.basicblocks = 0;
1795 tso->gran.allocs = 0;
1796 tso->gran.exectime = 0;
1797 tso->gran.fetchtime = 0;
1798 tso->gran.fetchcount = 0;
1799 tso->gran.blocktime = 0;
1800 tso->gran.blockcount = 0;
1801 tso->gran.blockedat = 0;
1802 tso->gran.globalsparks = 0;
1803 tso->gran.localsparks = 0;
1804 if (RtsFlags.GranFlags.Light)
1805 tso->gran.clock = Now; /* local clock */
1807 tso->gran.clock = 0;
1809 IF_DEBUG(gran,printTSO(tso));
1812 tso->par.magic = TSO_MAGIC; // debugging only
1814 tso->par.sparkname = 0;
1815 tso->par.startedat = CURRENT_TIME;
1816 tso->par.exported = 0;
1817 tso->par.basicblocks = 0;
1818 tso->par.allocs = 0;
1819 tso->par.exectime = 0;
1820 tso->par.fetchtime = 0;
1821 tso->par.fetchcount = 0;
1822 tso->par.blocktime = 0;
1823 tso->par.blockcount = 0;
1824 tso->par.blockedat = 0;
1825 tso->par.globalsparks = 0;
1826 tso->par.localsparks = 0;
1830 globalGranStats.tot_threads_created++;
1831 globalGranStats.threads_created_on_PE[CurrentProc]++;
1832 globalGranStats.tot_sq_len += spark_queue_len(CurrentProc);
1833 globalGranStats.tot_sq_probes++;
1835 // collect parallel global statistics (currently done together with GC stats)
1836 if (RtsFlags.ParFlags.ParStats.Global &&
1837 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
1838 //fprintf(stderr, "Creating thread %d @ %11.2f\n", tso->id, usertime());
1839 globalParStats.tot_threads_created++;
1845 belch("==__ schedule: Created TSO %d (%p);",
1846 CurrentProc, tso, tso->id));
1848 IF_PAR_DEBUG(verbose,
1849 belch("==__ schedule: Created TSO %d (%p); %d threads active",
1850 tso->id, tso, advisory_thread_count));
1852 IF_DEBUG(scheduler,sched_belch("created thread %ld, stack size = %lx words",
1853 tso->id, tso->stack_size));
1860 all parallel thread creation calls should fall through the following routine.
1863 createSparkThread(rtsSpark spark)
1865 ASSERT(spark != (rtsSpark)NULL);
1866 if (advisory_thread_count >= RtsFlags.ParFlags.maxThreads)
1868 barf("{createSparkThread}Daq ghuH: refusing to create another thread; no more than %d threads allowed (currently %d)",
1869 RtsFlags.ParFlags.maxThreads, advisory_thread_count);
1870 return END_TSO_QUEUE;
1874 tso = createThread(RtsFlags.GcFlags.initialStkSize);
1875 if (tso==END_TSO_QUEUE)
1876 barf("createSparkThread: Cannot create TSO");
1878 tso->priority = AdvisoryPriority;
1880 pushClosure(tso,spark);
1881 PUSH_ON_RUN_QUEUE(tso);
1882 advisory_thread_count++;
1889 Turn a spark into a thread.
1890 ToDo: fix for SMP (needs to acquire SCHED_MUTEX!)
1893 //@cindex activateSpark
1895 activateSpark (rtsSpark spark)
1899 tso = createSparkThread(spark);
1900 if (RtsFlags.ParFlags.ParStats.Full) {
1901 //ASSERT(run_queue_hd == END_TSO_QUEUE); // I think ...
1902 IF_PAR_DEBUG(verbose,
1903 belch("==^^ activateSpark: turning spark of closure %p (%s) into a thread",
1904 (StgClosure *)spark, info_type((StgClosure *)spark)));
1906 // ToDo: fwd info on local/global spark to thread -- HWL
1907 // tso->gran.exported = spark->exported;
1908 // tso->gran.locked = !spark->global;
1909 // tso->gran.sparkname = spark->name;
1915 static SchedulerStatus waitThread_(/*out*/StgMainThread* m
1916 #if defined(THREADED_RTS)
1917 , rtsBool blockWaiting
1922 /* ---------------------------------------------------------------------------
1925 * scheduleThread puts a thread on the head of the runnable queue.
1926 * This will usually be done immediately after a thread is created.
1927 * The caller of scheduleThread must create the thread using e.g.
1928 * createThread and push an appropriate closure
1929 * on this thread's stack before the scheduler is invoked.
1930 * ------------------------------------------------------------------------ */
1932 static void scheduleThread_ (StgTSO* tso, rtsBool createTask);
1935 scheduleThread_(StgTSO *tso
1936 , rtsBool createTask
1937 #if !defined(THREADED_RTS)
1942 ACQUIRE_LOCK(&sched_mutex);
1944 /* Put the new thread on the head of the runnable queue. The caller
1945 * better push an appropriate closure on this thread's stack
1946 * beforehand. In the SMP case, the thread may start running as
1947 * soon as we release the scheduler lock below.
1949 PUSH_ON_RUN_QUEUE(tso);
1950 #if defined(THREADED_RTS)
1951 /* If main() is scheduling a thread, don't bother creating a
1955 startTask(taskStart);
1961 IF_DEBUG(scheduler,printTSO(tso));
1963 RELEASE_LOCK(&sched_mutex);
1966 void scheduleThread(StgTSO* tso)
1968 scheduleThread_(tso, rtsFalse);
1972 scheduleWaitThread(StgTSO* tso, /*[out]*/HaskellObj* ret)
1976 m = stgMallocBytes(sizeof(StgMainThread), "waitThread");
1980 #if defined(RTS_SUPPORTS_THREADS)
1981 initCondition(&m->wakeup);
1984 /* Put the thread on the main-threads list prior to scheduling the TSO.
1985 Failure to do so introduces a race condition in the MT case (as
1986 identified by Wolfgang Thaller), whereby the new task/OS thread
1987 created by scheduleThread_() would complete prior to the thread
1988 that spawned it managed to put 'itself' on the main-threads list.
1989 The upshot of it all being that the worker thread wouldn't get to
1990 signal the completion of the its work item for the main thread to
1991 see (==> it got stuck waiting.) -- sof 6/02.
1993 ACQUIRE_LOCK(&sched_mutex);
1994 IF_DEBUG(scheduler, sched_belch("== scheduler: waiting for thread (%d)\n", tso->id));
1996 m->link = main_threads;
1999 /* Inefficient (scheduleThread_() acquires it again right away),
2000 * but obviously correct.
2002 RELEASE_LOCK(&sched_mutex);
2004 scheduleThread_(tso, rtsTrue);
2005 #if defined(THREADED_RTS)
2006 return waitThread_(m, rtsTrue);
2008 return waitThread_(m);
2012 /* ---------------------------------------------------------------------------
2015 * Initialise the scheduler. This resets all the queues - if the
2016 * queues contained any threads, they'll be garbage collected at the
2019 * ------------------------------------------------------------------------ */
2023 term_handler(int sig STG_UNUSED)
2026 ACQUIRE_LOCK(&term_mutex);
2028 RELEASE_LOCK(&term_mutex);
2039 for (i=0; i<=MAX_PROC; i++) {
2040 run_queue_hds[i] = END_TSO_QUEUE;
2041 run_queue_tls[i] = END_TSO_QUEUE;
2042 blocked_queue_hds[i] = END_TSO_QUEUE;
2043 blocked_queue_tls[i] = END_TSO_QUEUE;
2044 ccalling_threadss[i] = END_TSO_QUEUE;
2045 sleeping_queue = END_TSO_QUEUE;
2048 run_queue_hd = END_TSO_QUEUE;
2049 run_queue_tl = END_TSO_QUEUE;
2050 blocked_queue_hd = END_TSO_QUEUE;
2051 blocked_queue_tl = END_TSO_QUEUE;
2052 sleeping_queue = END_TSO_QUEUE;
2055 suspended_ccalling_threads = END_TSO_QUEUE;
2057 main_threads = NULL;
2058 all_threads = END_TSO_QUEUE;
2063 RtsFlags.ConcFlags.ctxtSwitchTicks =
2064 RtsFlags.ConcFlags.ctxtSwitchTime / TICK_MILLISECS;
2066 #if defined(RTS_SUPPORTS_THREADS)
2067 /* Initialise the mutex and condition variables used by
2069 initMutex(&sched_mutex);
2070 initMutex(&term_mutex);
2071 initMutex(&thread_id_mutex);
2073 initCondition(&thread_ready_cond);
2077 initCondition(&gc_pending_cond);
2080 #if defined(RTS_SUPPORTS_THREADS)
2081 ACQUIRE_LOCK(&sched_mutex);
2084 /* Install the SIGHUP handler */
2087 struct sigaction action,oact;
2089 action.sa_handler = term_handler;
2090 sigemptyset(&action.sa_mask);
2091 action.sa_flags = 0;
2092 if (sigaction(SIGTERM, &action, &oact) != 0) {
2093 barf("can't install TERM handler");
2098 /* A capability holds the state a native thread needs in
2099 * order to execute STG code. At least one capability is
2100 * floating around (only SMP builds have more than one).
2104 #if defined(RTS_SUPPORTS_THREADS)
2105 /* start our haskell execution tasks */
2107 startTaskManager(RtsFlags.ParFlags.nNodes, taskStart);
2109 startTaskManager(0,taskStart);
2113 #if /* defined(SMP) ||*/ defined(PAR)
2117 #if defined(RTS_SUPPORTS_THREADS)
2118 RELEASE_LOCK(&sched_mutex);
2124 exitScheduler( void )
2126 #if defined(RTS_SUPPORTS_THREADS)
2129 shutting_down_scheduler = rtsTrue;
2132 /* -----------------------------------------------------------------------------
2133 Managing the per-task allocation areas.
2135 Each capability comes with an allocation area. These are
2136 fixed-length block lists into which allocation can be done.
2138 ToDo: no support for two-space collection at the moment???
2139 -------------------------------------------------------------------------- */
2141 /* -----------------------------------------------------------------------------
2142 * waitThread is the external interface for running a new computation
2143 * and waiting for the result.
2145 * In the non-SMP case, we create a new main thread, push it on the
2146 * main-thread stack, and invoke the scheduler to run it. The
2147 * scheduler will return when the top main thread on the stack has
2148 * completed or died, and fill in the necessary fields of the
2149 * main_thread structure.
2151 * In the SMP case, we create a main thread as before, but we then
2152 * create a new condition variable and sleep on it. When our new
2153 * main thread has completed, we'll be woken up and the status/result
2154 * will be in the main_thread struct.
2155 * -------------------------------------------------------------------------- */
2158 howManyThreadsAvail ( void )
2162 for (q = run_queue_hd; q != END_TSO_QUEUE; q = q->link)
2164 for (q = blocked_queue_hd; q != END_TSO_QUEUE; q = q->link)
2166 for (q = sleeping_queue; q != END_TSO_QUEUE; q = q->link)
2172 finishAllThreads ( void )
2175 while (run_queue_hd != END_TSO_QUEUE) {
2176 waitThread ( run_queue_hd, NULL);
2178 while (blocked_queue_hd != END_TSO_QUEUE) {
2179 waitThread ( blocked_queue_hd, NULL);
2181 while (sleeping_queue != END_TSO_QUEUE) {
2182 waitThread ( blocked_queue_hd, NULL);
2185 (blocked_queue_hd != END_TSO_QUEUE ||
2186 run_queue_hd != END_TSO_QUEUE ||
2187 sleeping_queue != END_TSO_QUEUE);
2191 waitThread(StgTSO *tso, /*out*/StgClosure **ret)
2195 m = stgMallocBytes(sizeof(StgMainThread), "waitThread");
2199 #if defined(RTS_SUPPORTS_THREADS)
2200 initCondition(&m->wakeup);
2203 /* see scheduleWaitThread() comment */
2204 ACQUIRE_LOCK(&sched_mutex);
2205 IF_DEBUG(scheduler, sched_belch("== scheduler: waiting for thread (%d)\n", tso->id));
2206 m->link = main_threads;
2208 RELEASE_LOCK(&sched_mutex);
2210 IF_DEBUG(scheduler, sched_belch("== scheduler: waiting for thread (%d)\n", tso->id));
2211 #if defined(THREADED_RTS)
2212 return waitThread_(m, rtsFalse);
2214 return waitThread_(m);
2220 waitThread_(StgMainThread* m
2221 #if defined(THREADED_RTS)
2222 , rtsBool blockWaiting
2226 SchedulerStatus stat;
2228 IF_DEBUG(scheduler, sched_belch("== scheduler: new main thread (%d)\n", m->tso->id));
2230 #if defined(RTS_SUPPORTS_THREADS)
2232 # if defined(THREADED_RTS)
2233 if (!blockWaiting) {
2234 /* In the threaded case, the OS thread that called main()
2235 * gets to enter the RTS directly without going via another
2239 ASSERT(m->stat != NoStatus);
2243 ACQUIRE_LOCK(&sched_mutex);
2245 waitCondition(&m->wakeup, &sched_mutex);
2246 } while (m->stat == NoStatus);
2249 /* GranSim specific init */
2250 CurrentTSO = m->tso; // the TSO to run
2251 procStatus[MainProc] = Busy; // status of main PE
2252 CurrentProc = MainProc; // PE to run it on
2256 RELEASE_LOCK(&sched_mutex);
2258 ASSERT(m->stat != NoStatus);
2263 #if defined(RTS_SUPPORTS_THREADS)
2264 closeCondition(&m->wakeup);
2267 IF_DEBUG(scheduler, fprintf(stderr, "== scheduler: main thread (%d) finished\n",
2271 #if defined(THREADED_RTS)
2274 RELEASE_LOCK(&sched_mutex);
2279 //@node Run queue code, Garbage Collextion Routines, Suspend and Resume, Main scheduling code
2280 //@subsection Run queue code
2284 NB: In GranSim we have many run queues; run_queue_hd is actually a macro
2285 unfolding to run_queue_hds[CurrentProc], thus CurrentProc is an
2286 implicit global variable that has to be correct when calling these
2290 /* Put the new thread on the head of the runnable queue.
2291 * The caller of createThread better push an appropriate closure
2292 * on this thread's stack before the scheduler is invoked.
2294 static /* inline */ void
2295 add_to_run_queue(tso)
2298 ASSERT(tso!=run_queue_hd && tso!=run_queue_tl);
2299 tso->link = run_queue_hd;
2301 if (run_queue_tl == END_TSO_QUEUE) {
2306 /* Put the new thread at the end of the runnable queue. */
2307 static /* inline */ void
2308 push_on_run_queue(tso)
2311 ASSERT(get_itbl((StgClosure *)tso)->type == TSO);
2312 ASSERT(run_queue_hd!=NULL && run_queue_tl!=NULL);
2313 ASSERT(tso!=run_queue_hd && tso!=run_queue_tl);
2314 if (run_queue_hd == END_TSO_QUEUE) {
2317 run_queue_tl->link = tso;
2323 Should be inlined because it's used very often in schedule. The tso
2324 argument is actually only needed in GranSim, where we want to have the
2325 possibility to schedule *any* TSO on the run queue, irrespective of the
2326 actual ordering. Therefore, if tso is not the nil TSO then we traverse
2327 the run queue and dequeue the tso, adjusting the links in the queue.
2329 //@cindex take_off_run_queue
2330 static /* inline */ StgTSO*
2331 take_off_run_queue(StgTSO *tso) {
2335 qetlaHbogh Qu' ngaSbogh ghomDaQ {tso} yIteq!
2337 if tso is specified, unlink that tso from the run_queue (doesn't have
2338 to be at the beginning of the queue); GranSim only
2340 if (tso!=END_TSO_QUEUE) {
2341 /* find tso in queue */
2342 for (t=run_queue_hd, prev=END_TSO_QUEUE;
2343 t!=END_TSO_QUEUE && t!=tso;
2347 /* now actually dequeue the tso */
2348 if (prev!=END_TSO_QUEUE) {
2349 ASSERT(run_queue_hd!=t);
2350 prev->link = t->link;
2352 /* t is at beginning of thread queue */
2353 ASSERT(run_queue_hd==t);
2354 run_queue_hd = t->link;
2356 /* t is at end of thread queue */
2357 if (t->link==END_TSO_QUEUE) {
2358 ASSERT(t==run_queue_tl);
2359 run_queue_tl = prev;
2361 ASSERT(run_queue_tl!=t);
2363 t->link = END_TSO_QUEUE;
2365 /* take tso from the beginning of the queue; std concurrent code */
2367 if (t != END_TSO_QUEUE) {
2368 run_queue_hd = t->link;
2369 t->link = END_TSO_QUEUE;
2370 if (run_queue_hd == END_TSO_QUEUE) {
2371 run_queue_tl = END_TSO_QUEUE;
2380 //@node Garbage Collextion Routines, Blocking Queue Routines, Run queue code, Main scheduling code
2381 //@subsection Garbage Collextion Routines
2383 /* ---------------------------------------------------------------------------
2384 Where are the roots that we know about?
2386 - all the threads on the runnable queue
2387 - all the threads on the blocked queue
2388 - all the threads on the sleeping queue
2389 - all the thread currently executing a _ccall_GC
2390 - all the "main threads"
2392 ------------------------------------------------------------------------ */
2394 /* This has to be protected either by the scheduler monitor, or by the
2395 garbage collection monitor (probably the latter).
2400 GetRoots(evac_fn evac)
2405 for (i=0; i<=RtsFlags.GranFlags.proc; i++) {
2406 if ((run_queue_hds[i] != END_TSO_QUEUE) && ((run_queue_hds[i] != NULL)))
2407 evac((StgClosure **)&run_queue_hds[i]);
2408 if ((run_queue_tls[i] != END_TSO_QUEUE) && ((run_queue_tls[i] != NULL)))
2409 evac((StgClosure **)&run_queue_tls[i]);
2411 if ((blocked_queue_hds[i] != END_TSO_QUEUE) && ((blocked_queue_hds[i] != NULL)))
2412 evac((StgClosure **)&blocked_queue_hds[i]);
2413 if ((blocked_queue_tls[i] != END_TSO_QUEUE) && ((blocked_queue_tls[i] != NULL)))
2414 evac((StgClosure **)&blocked_queue_tls[i]);
2415 if ((ccalling_threadss[i] != END_TSO_QUEUE) && ((ccalling_threadss[i] != NULL)))
2416 evac((StgClosure **)&ccalling_threads[i]);
2423 if (run_queue_hd != END_TSO_QUEUE) {
2424 ASSERT(run_queue_tl != END_TSO_QUEUE);
2425 evac((StgClosure **)&run_queue_hd);
2426 evac((StgClosure **)&run_queue_tl);
2429 if (blocked_queue_hd != END_TSO_QUEUE) {
2430 ASSERT(blocked_queue_tl != END_TSO_QUEUE);
2431 evac((StgClosure **)&blocked_queue_hd);
2432 evac((StgClosure **)&blocked_queue_tl);
2435 if (sleeping_queue != END_TSO_QUEUE) {
2436 evac((StgClosure **)&sleeping_queue);
2440 if (suspended_ccalling_threads != END_TSO_QUEUE) {
2441 evac((StgClosure **)&suspended_ccalling_threads);
2444 #if defined(PAR) || defined(GRAN)
2445 markSparkQueue(evac);
2449 /* -----------------------------------------------------------------------------
2452 This is the interface to the garbage collector from Haskell land.
2453 We provide this so that external C code can allocate and garbage
2454 collect when called from Haskell via _ccall_GC.
2456 It might be useful to provide an interface whereby the programmer
2457 can specify more roots (ToDo).
2459 This needs to be protected by the GC condition variable above. KH.
2460 -------------------------------------------------------------------------- */
2462 void (*extra_roots)(evac_fn);
2467 /* Obligated to hold this lock upon entry */
2468 ACQUIRE_LOCK(&sched_mutex);
2469 GarbageCollect(GetRoots,rtsFalse);
2470 RELEASE_LOCK(&sched_mutex);
2474 performMajorGC(void)
2476 ACQUIRE_LOCK(&sched_mutex);
2477 GarbageCollect(GetRoots,rtsTrue);
2478 RELEASE_LOCK(&sched_mutex);
2482 AllRoots(evac_fn evac)
2484 GetRoots(evac); // the scheduler's roots
2485 extra_roots(evac); // the user's roots
2489 performGCWithRoots(void (*get_roots)(evac_fn))
2491 ACQUIRE_LOCK(&sched_mutex);
2492 extra_roots = get_roots;
2493 GarbageCollect(AllRoots,rtsFalse);
2494 RELEASE_LOCK(&sched_mutex);
2497 /* -----------------------------------------------------------------------------
2500 If the thread has reached its maximum stack size, then raise the
2501 StackOverflow exception in the offending thread. Otherwise
2502 relocate the TSO into a larger chunk of memory and adjust its stack
2504 -------------------------------------------------------------------------- */
2507 threadStackOverflow(StgTSO *tso)
2509 nat new_stack_size, new_tso_size, diff, stack_words;
2513 IF_DEBUG(sanity,checkTSO(tso));
2514 if (tso->stack_size >= tso->max_stack_size) {
2517 belch("@@ threadStackOverflow of TSO %d (%p): stack too large (now %ld; max is %ld",
2518 tso->id, tso, tso->stack_size, tso->max_stack_size);
2519 /* If we're debugging, just print out the top of the stack */
2520 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2523 /* Send this thread the StackOverflow exception */
2524 raiseAsync(tso, (StgClosure *)stackOverflow_closure);
2528 /* Try to double the current stack size. If that takes us over the
2529 * maximum stack size for this thread, then use the maximum instead.
2530 * Finally round up so the TSO ends up as a whole number of blocks.
2532 new_stack_size = stg_min(tso->stack_size * 2, tso->max_stack_size);
2533 new_tso_size = (nat)BLOCK_ROUND_UP(new_stack_size * sizeof(W_) +
2534 TSO_STRUCT_SIZE)/sizeof(W_);
2535 new_tso_size = round_to_mblocks(new_tso_size); /* Be MBLOCK-friendly */
2536 new_stack_size = new_tso_size - TSO_STRUCT_SIZEW;
2538 IF_DEBUG(scheduler, fprintf(stderr,"== scheduler: increasing stack size from %d words to %d.\n", tso->stack_size, new_stack_size));
2540 dest = (StgTSO *)allocate(new_tso_size);
2541 TICK_ALLOC_TSO(new_stack_size,0);
2543 /* copy the TSO block and the old stack into the new area */
2544 memcpy(dest,tso,TSO_STRUCT_SIZE);
2545 stack_words = tso->stack + tso->stack_size - tso->sp;
2546 new_sp = (P_)dest + new_tso_size - stack_words;
2547 memcpy(new_sp, tso->sp, stack_words * sizeof(W_));
2549 /* relocate the stack pointers... */
2550 diff = (P_)new_sp - (P_)tso->sp; /* In *words* */
2551 dest->su = (StgUpdateFrame *) ((P_)dest->su + diff);
2553 dest->stack_size = new_stack_size;
2555 /* and relocate the update frame list */
2556 relocate_stack(dest, diff);
2558 /* Mark the old TSO as relocated. We have to check for relocated
2559 * TSOs in the garbage collector and any primops that deal with TSOs.
2561 * It's important to set the sp and su values to just beyond the end
2562 * of the stack, so we don't attempt to scavenge any part of the
2565 tso->what_next = ThreadRelocated;
2567 tso->sp = (P_)&(tso->stack[tso->stack_size]);
2568 tso->su = (StgUpdateFrame *)tso->sp;
2569 tso->why_blocked = NotBlocked;
2570 dest->mut_link = NULL;
2572 IF_PAR_DEBUG(verbose,
2573 belch("@@ threadStackOverflow of TSO %d (now at %p): stack size increased to %ld",
2574 tso->id, tso, tso->stack_size);
2575 /* If we're debugging, just print out the top of the stack */
2576 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2579 IF_DEBUG(sanity,checkTSO(tso));
2581 IF_DEBUG(scheduler,printTSO(dest));
2587 //@node Blocking Queue Routines, Exception Handling Routines, Garbage Collextion Routines, Main scheduling code
2588 //@subsection Blocking Queue Routines
2590 /* ---------------------------------------------------------------------------
2591 Wake up a queue that was blocked on some resource.
2592 ------------------------------------------------------------------------ */
2596 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2601 unblockCount ( StgBlockingQueueElement *bqe, StgClosure *node )
2603 /* write RESUME events to log file and
2604 update blocked and fetch time (depending on type of the orig closure) */
2605 if (RtsFlags.ParFlags.ParStats.Full) {
2606 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
2607 GR_RESUMEQ, ((StgTSO *)bqe), ((StgTSO *)bqe)->block_info.closure,
2608 0, 0 /* spark_queue_len(ADVISORY_POOL) */);
2609 if (EMPTY_RUN_QUEUE())
2610 emitSchedule = rtsTrue;
2612 switch (get_itbl(node)->type) {
2614 ((StgTSO *)bqe)->par.fetchtime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2619 ((StgTSO *)bqe)->par.blocktime += CURRENT_TIME-((StgTSO *)bqe)->par.blockedat;
2626 barf("{unblockOneLocked}Daq Qagh: unexpected closure in blocking queue");
2633 static StgBlockingQueueElement *
2634 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2637 PEs node_loc, tso_loc;
2639 node_loc = where_is(node); // should be lifted out of loop
2640 tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2641 tso_loc = where_is((StgClosure *)tso);
2642 if (IS_LOCAL_TO(PROCS(node),tso_loc)) { // TSO is local
2643 /* !fake_fetch => TSO is on CurrentProc is same as IS_LOCAL_TO */
2644 ASSERT(CurrentProc!=node_loc || tso_loc==CurrentProc);
2645 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.lunblocktime;
2646 // insertThread(tso, node_loc);
2647 new_event(tso_loc, tso_loc, CurrentTime[CurrentProc],
2649 tso, node, (rtsSpark*)NULL);
2650 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2653 } else { // TSO is remote (actually should be FMBQ)
2654 CurrentTime[CurrentProc] += RtsFlags.GranFlags.Costs.mpacktime +
2655 RtsFlags.GranFlags.Costs.gunblocktime +
2656 RtsFlags.GranFlags.Costs.latency;
2657 new_event(tso_loc, CurrentProc, CurrentTime[CurrentProc],
2659 tso, node, (rtsSpark*)NULL);
2660 tso->link = END_TSO_QUEUE; // overwrite link just to be sure
2663 /* the thread-queue-overhead is accounted for in either Resume or UnblockThread */
2665 fprintf(stderr," %s TSO %d (%p) [PE %d] (block_info.closure=%p) (next=%p) ,",
2666 (node_loc==tso_loc ? "Local" : "Global"),
2667 tso->id, tso, CurrentProc, tso->block_info.closure, tso->link));
2668 tso->block_info.closure = NULL;
2669 IF_DEBUG(scheduler,belch("-- Waking up thread %ld (%p)",
2673 static StgBlockingQueueElement *
2674 unblockOneLocked(StgBlockingQueueElement *bqe, StgClosure *node)
2676 StgBlockingQueueElement *next;
2678 switch (get_itbl(bqe)->type) {
2680 ASSERT(((StgTSO *)bqe)->why_blocked != NotBlocked);
2681 /* if it's a TSO just push it onto the run_queue */
2683 // ((StgTSO *)bqe)->link = END_TSO_QUEUE; // debugging?
2684 PUSH_ON_RUN_QUEUE((StgTSO *)bqe);
2686 unblockCount(bqe, node);
2687 /* reset blocking status after dumping event */
2688 ((StgTSO *)bqe)->why_blocked = NotBlocked;
2692 /* if it's a BLOCKED_FETCH put it on the PendingFetches list */
2694 bqe->link = (StgBlockingQueueElement *)PendingFetches;
2695 PendingFetches = (StgBlockedFetch *)bqe;
2699 /* can ignore this case in a non-debugging setup;
2700 see comments on RBHSave closures above */
2702 /* check that the closure is an RBHSave closure */
2703 ASSERT(get_itbl((StgClosure *)bqe) == &stg_RBH_Save_0_info ||
2704 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_1_info ||
2705 get_itbl((StgClosure *)bqe) == &stg_RBH_Save_2_info);
2709 barf("{unblockOneLocked}Daq Qagh: Unexpected IP (%#lx; %s) in blocking queue at %#lx\n",
2710 get_itbl((StgClosure *)bqe), info_type((StgClosure *)bqe),
2714 IF_PAR_DEBUG(bq, fprintf(stderr, ", %p (%s)", bqe, info_type((StgClosure*)bqe)));
2718 #else /* !GRAN && !PAR */
2720 unblockOneLocked(StgTSO *tso)
2724 ASSERT(get_itbl(tso)->type == TSO);
2725 ASSERT(tso->why_blocked != NotBlocked);
2726 tso->why_blocked = NotBlocked;
2728 PUSH_ON_RUN_QUEUE(tso);
2730 IF_DEBUG(scheduler,sched_belch("waking up thread %ld", tso->id));
2735 #if defined(GRAN) || defined(PAR)
2736 inline StgBlockingQueueElement *
2737 unblockOne(StgBlockingQueueElement *bqe, StgClosure *node)
2739 ACQUIRE_LOCK(&sched_mutex);
2740 bqe = unblockOneLocked(bqe, node);
2741 RELEASE_LOCK(&sched_mutex);
2746 unblockOne(StgTSO *tso)
2748 ACQUIRE_LOCK(&sched_mutex);
2749 tso = unblockOneLocked(tso);
2750 RELEASE_LOCK(&sched_mutex);
2757 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2759 StgBlockingQueueElement *bqe;
2764 belch("##-_ AwBQ for node %p on PE %d @ %ld by TSO %d (%p): ", \
2765 node, CurrentProc, CurrentTime[CurrentProc],
2766 CurrentTSO->id, CurrentTSO));
2768 node_loc = where_is(node);
2770 ASSERT(q == END_BQ_QUEUE ||
2771 get_itbl(q)->type == TSO || // q is either a TSO or an RBHSave
2772 get_itbl(q)->type == CONSTR); // closure (type constructor)
2773 ASSERT(is_unique(node));
2775 /* FAKE FETCH: magically copy the node to the tso's proc;
2776 no Fetch necessary because in reality the node should not have been
2777 moved to the other PE in the first place
2779 if (CurrentProc!=node_loc) {
2781 belch("## node %p is on PE %d but CurrentProc is %d (TSO %d); assuming fake fetch and adjusting bitmask (old: %#x)",
2782 node, node_loc, CurrentProc, CurrentTSO->id,
2783 // CurrentTSO, where_is(CurrentTSO),
2784 node->header.gran.procs));
2785 node->header.gran.procs = (node->header.gran.procs) | PE_NUMBER(CurrentProc);
2787 belch("## new bitmask of node %p is %#x",
2788 node, node->header.gran.procs));
2789 if (RtsFlags.GranFlags.GranSimStats.Global) {
2790 globalGranStats.tot_fake_fetches++;
2795 // ToDo: check: ASSERT(CurrentProc==node_loc);
2796 while (get_itbl(bqe)->type==TSO) { // q != END_TSO_QUEUE) {
2799 bqe points to the current element in the queue
2800 next points to the next element in the queue
2802 //tso = (StgTSO *)bqe; // wastes an assignment to get the type right
2803 //tso_loc = where_is(tso);
2805 bqe = unblockOneLocked(bqe, node);
2808 /* if this is the BQ of an RBH, we have to put back the info ripped out of
2809 the closure to make room for the anchor of the BQ */
2810 if (bqe!=END_BQ_QUEUE) {
2811 ASSERT(get_itbl(node)->type == RBH && get_itbl(bqe)->type == CONSTR);
2813 ASSERT((info_ptr==&RBH_Save_0_info) ||
2814 (info_ptr==&RBH_Save_1_info) ||
2815 (info_ptr==&RBH_Save_2_info));
2817 /* cf. convertToRBH in RBH.c for writing the RBHSave closure */
2818 ((StgRBH *)node)->blocking_queue = (StgBlockingQueueElement *)((StgRBHSave *)bqe)->payload[0];
2819 ((StgRBH *)node)->mut_link = (StgMutClosure *)((StgRBHSave *)bqe)->payload[1];
2822 belch("## Filled in RBH_Save for %p (%s) at end of AwBQ",
2823 node, info_type(node)));
2826 /* statistics gathering */
2827 if (RtsFlags.GranFlags.GranSimStats.Global) {
2828 // globalGranStats.tot_bq_processing_time += bq_processing_time;
2829 globalGranStats.tot_bq_len += len; // total length of all bqs awakened
2830 // globalGranStats.tot_bq_len_local += len_local; // same for local TSOs only
2831 globalGranStats.tot_awbq++; // total no. of bqs awakened
2834 fprintf(stderr,"## BQ Stats of %p: [%d entries] %s\n",
2835 node, len, (bqe!=END_BQ_QUEUE) ? "RBH" : ""));
2839 awakenBlockedQueue(StgBlockingQueueElement *q, StgClosure *node)
2841 StgBlockingQueueElement *bqe;
2843 ACQUIRE_LOCK(&sched_mutex);
2845 IF_PAR_DEBUG(verbose,
2846 belch("##-_ AwBQ for node %p on [%x]: ",
2850 if(get_itbl(q)->type == CONSTR || q==END_BQ_QUEUE) {
2851 IF_PAR_DEBUG(verbose, belch("## ... nothing to unblock so lets just return. RFP (BUG?)"));
2856 ASSERT(q == END_BQ_QUEUE ||
2857 get_itbl(q)->type == TSO ||
2858 get_itbl(q)->type == BLOCKED_FETCH ||
2859 get_itbl(q)->type == CONSTR);
2862 while (get_itbl(bqe)->type==TSO ||
2863 get_itbl(bqe)->type==BLOCKED_FETCH) {
2864 bqe = unblockOneLocked(bqe, node);
2866 RELEASE_LOCK(&sched_mutex);
2869 #else /* !GRAN && !PAR */
2871 awakenBlockedQueue(StgTSO *tso)
2873 ACQUIRE_LOCK(&sched_mutex);
2874 while (tso != END_TSO_QUEUE) {
2875 tso = unblockOneLocked(tso);
2877 RELEASE_LOCK(&sched_mutex);
2881 //@node Exception Handling Routines, Debugging Routines, Blocking Queue Routines, Main scheduling code
2882 //@subsection Exception Handling Routines
2884 /* ---------------------------------------------------------------------------
2886 - usually called inside a signal handler so it mustn't do anything fancy.
2887 ------------------------------------------------------------------------ */
2890 interruptStgRts(void)
2896 /* -----------------------------------------------------------------------------
2899 This is for use when we raise an exception in another thread, which
2901 This has nothing to do with the UnblockThread event in GranSim. -- HWL
2902 -------------------------------------------------------------------------- */
2904 #if defined(GRAN) || defined(PAR)
2906 NB: only the type of the blocking queue is different in GranSim and GUM
2907 the operations on the queue-elements are the same
2908 long live polymorphism!
2910 Locks: sched_mutex is held upon entry and exit.
2914 unblockThread(StgTSO *tso)
2916 StgBlockingQueueElement *t, **last;
2918 switch (tso->why_blocked) {
2921 return; /* not blocked */
2924 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
2926 StgBlockingQueueElement *last_tso = END_BQ_QUEUE;
2927 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
2929 last = (StgBlockingQueueElement **)&mvar->head;
2930 for (t = (StgBlockingQueueElement *)mvar->head;
2932 last = &t->link, last_tso = t, t = t->link) {
2933 if (t == (StgBlockingQueueElement *)tso) {
2934 *last = (StgBlockingQueueElement *)tso->link;
2935 if (mvar->tail == tso) {
2936 mvar->tail = (StgTSO *)last_tso;
2941 barf("unblockThread (MVAR): TSO not found");
2944 case BlockedOnBlackHole:
2945 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
2947 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
2949 last = &bq->blocking_queue;
2950 for (t = bq->blocking_queue;
2952 last = &t->link, t = t->link) {
2953 if (t == (StgBlockingQueueElement *)tso) {
2954 *last = (StgBlockingQueueElement *)tso->link;
2958 barf("unblockThread (BLACKHOLE): TSO not found");
2961 case BlockedOnException:
2963 StgTSO *target = tso->block_info.tso;
2965 ASSERT(get_itbl(target)->type == TSO);
2967 if (target->what_next == ThreadRelocated) {
2968 target = target->link;
2969 ASSERT(get_itbl(target)->type == TSO);
2972 ASSERT(target->blocked_exceptions != NULL);
2974 last = (StgBlockingQueueElement **)&target->blocked_exceptions;
2975 for (t = (StgBlockingQueueElement *)target->blocked_exceptions;
2977 last = &t->link, t = t->link) {
2978 ASSERT(get_itbl(t)->type == TSO);
2979 if (t == (StgBlockingQueueElement *)tso) {
2980 *last = (StgBlockingQueueElement *)tso->link;
2984 barf("unblockThread (Exception): TSO not found");
2988 case BlockedOnWrite:
2990 /* take TSO off blocked_queue */
2991 StgBlockingQueueElement *prev = NULL;
2992 for (t = (StgBlockingQueueElement *)blocked_queue_hd; t != END_BQ_QUEUE;
2993 prev = t, t = t->link) {
2994 if (t == (StgBlockingQueueElement *)tso) {
2996 blocked_queue_hd = (StgTSO *)t->link;
2997 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
2998 blocked_queue_tl = END_TSO_QUEUE;
3001 prev->link = t->link;
3002 if ((StgBlockingQueueElement *)blocked_queue_tl == t) {
3003 blocked_queue_tl = (StgTSO *)prev;
3009 barf("unblockThread (I/O): TSO not found");
3012 case BlockedOnDelay:
3014 /* take TSO off sleeping_queue */
3015 StgBlockingQueueElement *prev = NULL;
3016 for (t = (StgBlockingQueueElement *)sleeping_queue; t != END_BQ_QUEUE;
3017 prev = t, t = t->link) {
3018 if (t == (StgBlockingQueueElement *)tso) {
3020 sleeping_queue = (StgTSO *)t->link;
3022 prev->link = t->link;
3027 barf("unblockThread (I/O): TSO not found");
3031 barf("unblockThread");
3035 tso->link = END_TSO_QUEUE;
3036 tso->why_blocked = NotBlocked;
3037 tso->block_info.closure = NULL;
3038 PUSH_ON_RUN_QUEUE(tso);
3042 unblockThread(StgTSO *tso)
3046 /* To avoid locking unnecessarily. */
3047 if (tso->why_blocked == NotBlocked) {
3051 switch (tso->why_blocked) {
3054 ASSERT(get_itbl(tso->block_info.closure)->type == MVAR);
3056 StgTSO *last_tso = END_TSO_QUEUE;
3057 StgMVar *mvar = (StgMVar *)(tso->block_info.closure);
3060 for (t = mvar->head; t != END_TSO_QUEUE;
3061 last = &t->link, last_tso = t, t = t->link) {
3064 if (mvar->tail == tso) {
3065 mvar->tail = last_tso;
3070 barf("unblockThread (MVAR): TSO not found");
3073 case BlockedOnBlackHole:
3074 ASSERT(get_itbl(tso->block_info.closure)->type == BLACKHOLE_BQ);
3076 StgBlockingQueue *bq = (StgBlockingQueue *)(tso->block_info.closure);
3078 last = &bq->blocking_queue;
3079 for (t = bq->blocking_queue; t != END_TSO_QUEUE;
3080 last = &t->link, t = t->link) {
3086 barf("unblockThread (BLACKHOLE): TSO not found");
3089 case BlockedOnException:
3091 StgTSO *target = tso->block_info.tso;
3093 ASSERT(get_itbl(target)->type == TSO);
3095 while (target->what_next == ThreadRelocated) {
3096 target = target->link;
3097 ASSERT(get_itbl(target)->type == TSO);
3100 ASSERT(target->blocked_exceptions != NULL);
3102 last = &target->blocked_exceptions;
3103 for (t = target->blocked_exceptions; t != END_TSO_QUEUE;
3104 last = &t->link, t = t->link) {
3105 ASSERT(get_itbl(t)->type == TSO);
3111 barf("unblockThread (Exception): TSO not found");
3115 case BlockedOnWrite:
3117 StgTSO *prev = NULL;
3118 for (t = blocked_queue_hd; t != END_TSO_QUEUE;
3119 prev = t, t = t->link) {
3122 blocked_queue_hd = t->link;
3123 if (blocked_queue_tl == t) {
3124 blocked_queue_tl = END_TSO_QUEUE;
3127 prev->link = t->link;
3128 if (blocked_queue_tl == t) {
3129 blocked_queue_tl = prev;
3135 barf("unblockThread (I/O): TSO not found");
3138 case BlockedOnDelay:
3140 StgTSO *prev = NULL;
3141 for (t = sleeping_queue; t != END_TSO_QUEUE;
3142 prev = t, t = t->link) {
3145 sleeping_queue = t->link;
3147 prev->link = t->link;
3152 barf("unblockThread (I/O): TSO not found");
3156 barf("unblockThread");
3160 tso->link = END_TSO_QUEUE;
3161 tso->why_blocked = NotBlocked;
3162 tso->block_info.closure = NULL;
3163 PUSH_ON_RUN_QUEUE(tso);
3167 /* -----------------------------------------------------------------------------
3170 * The following function implements the magic for raising an
3171 * asynchronous exception in an existing thread.
3173 * We first remove the thread from any queue on which it might be
3174 * blocked. The possible blockages are MVARs and BLACKHOLE_BQs.
3176 * We strip the stack down to the innermost CATCH_FRAME, building
3177 * thunks in the heap for all the active computations, so they can
3178 * be restarted if necessary. When we reach a CATCH_FRAME, we build
3179 * an application of the handler to the exception, and push it on
3180 * the top of the stack.
3182 * How exactly do we save all the active computations? We create an
3183 * AP_UPD for every UpdateFrame on the stack. Entering one of these
3184 * AP_UPDs pushes everything from the corresponding update frame
3185 * upwards onto the stack. (Actually, it pushes everything up to the
3186 * next update frame plus a pointer to the next AP_UPD object.
3187 * Entering the next AP_UPD object pushes more onto the stack until we
3188 * reach the last AP_UPD object - at which point the stack should look
3189 * exactly as it did when we killed the TSO and we can continue
3190 * execution by entering the closure on top of the stack.
3192 * We can also kill a thread entirely - this happens if either (a) the
3193 * exception passed to raiseAsync is NULL, or (b) there's no
3194 * CATCH_FRAME on the stack. In either case, we strip the entire
3195 * stack and replace the thread with a zombie.
3197 * Locks: sched_mutex held upon entry nor exit.
3199 * -------------------------------------------------------------------------- */
3202 deleteThread(StgTSO *tso)
3204 raiseAsync(tso,NULL);
3208 raiseAsyncWithLock(StgTSO *tso, StgClosure *exception)
3210 /* When raising async exs from contexts where sched_mutex isn't held;
3211 use raiseAsyncWithLock(). */
3212 ACQUIRE_LOCK(&sched_mutex);
3213 raiseAsync(tso,exception);
3214 RELEASE_LOCK(&sched_mutex);
3218 raiseAsync(StgTSO *tso, StgClosure *exception)
3220 StgUpdateFrame* su = tso->su;
3221 StgPtr sp = tso->sp;
3223 /* Thread already dead? */
3224 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
3228 IF_DEBUG(scheduler, sched_belch("raising exception in thread %ld.", tso->id));
3230 /* Remove it from any blocking queues */
3233 IF_DEBUG(scheduler, sched_belch("raising exception in thread %ld.", tso->id));
3234 /* The stack freezing code assumes there's a closure pointer on
3235 * the top of the stack. This isn't always the case with compiled
3236 * code, so we have to push a dummy closure on the top which just
3237 * returns to the next return address on the stack.
3239 if ( LOOKS_LIKE_GHC_INFO((void*)*sp) ) {
3240 *(--sp) = (W_)&stg_dummy_ret_closure;
3244 nat words = ((P_)su - (P_)sp) - 1;
3248 /* If we find a CATCH_FRAME, and we've got an exception to raise,
3249 * then build the THUNK raise(exception), and leave it on
3250 * top of the CATCH_FRAME ready to enter.
3252 if (get_itbl(su)->type == CATCH_FRAME && exception != NULL) {
3254 StgCatchFrame *cf = (StgCatchFrame *)su;
3258 /* we've got an exception to raise, so let's pass it to the
3259 * handler in this frame.
3261 raise = (StgClosure *)allocate(sizeofW(StgClosure)+1);
3262 TICK_ALLOC_SE_THK(1,0);
3263 SET_HDR(raise,&stg_raise_info,cf->header.prof.ccs);
3264 raise->payload[0] = exception;
3266 /* throw away the stack from Sp up to the CATCH_FRAME.
3270 /* Ensure that async excpetions are blocked now, so we don't get
3271 * a surprise exception before we get around to executing the
3274 if (tso->blocked_exceptions == NULL) {
3275 tso->blocked_exceptions = END_TSO_QUEUE;
3278 /* Put the newly-built THUNK on top of the stack, ready to execute
3279 * when the thread restarts.
3284 tso->what_next = ThreadEnterGHC;
3285 IF_DEBUG(sanity, checkTSO(tso));
3289 /* First build an AP_UPD consisting of the stack chunk above the
3290 * current update frame, with the top word on the stack as the
3293 ap = (StgAP_UPD *)allocate(AP_sizeW(words));
3298 ap->fun = (StgClosure *)sp[0];
3300 for(i=0; i < (nat)words; ++i) {
3301 ap->payload[i] = (StgClosure *)*sp++;
3304 switch (get_itbl(su)->type) {
3308 SET_HDR(ap,&stg_AP_UPD_info,su->header.prof.ccs /* ToDo */);
3309 TICK_ALLOC_UP_THK(words+1,0);
3312 fprintf(stderr, "scheduler: Updating ");
3313 printPtr((P_)su->updatee);
3314 fprintf(stderr, " with ");
3315 printObj((StgClosure *)ap);
3318 /* Replace the updatee with an indirection - happily
3319 * this will also wake up any threads currently
3320 * waiting on the result.
3322 * Warning: if we're in a loop, more than one update frame on
3323 * the stack may point to the same object. Be careful not to
3324 * overwrite an IND_OLDGEN in this case, because we'll screw
3325 * up the mutable lists. To be on the safe side, don't
3326 * overwrite any kind of indirection at all. See also
3327 * threadSqueezeStack in GC.c, where we have to make a similar
3330 if (!closure_IND(su->updatee)) {
3331 UPD_IND_NOLOCK(su->updatee,ap); /* revert the black hole */
3334 sp += sizeofW(StgUpdateFrame) -1;
3335 sp[0] = (W_)ap; /* push onto stack */
3341 StgCatchFrame *cf = (StgCatchFrame *)su;
3344 /* We want a PAP, not an AP_UPD. Fortunately, the
3345 * layout's the same.
3347 SET_HDR(ap,&stg_PAP_info,su->header.prof.ccs /* ToDo */);
3348 TICK_ALLOC_UPD_PAP(words+1,0);
3350 /* now build o = FUN(catch,ap,handler) */
3351 o = (StgClosure *)allocate(sizeofW(StgClosure)+2);
3352 TICK_ALLOC_FUN(2,0);
3353 SET_HDR(o,&stg_catch_info,su->header.prof.ccs /* ToDo */);
3354 o->payload[0] = (StgClosure *)ap;
3355 o->payload[1] = cf->handler;
3358 fprintf(stderr, "scheduler: Built ");
3359 printObj((StgClosure *)o);
3362 /* pop the old handler and put o on the stack */
3364 sp += sizeofW(StgCatchFrame) - 1;
3371 StgSeqFrame *sf = (StgSeqFrame *)su;
3374 SET_HDR(ap,&stg_PAP_info,su->header.prof.ccs /* ToDo */);
3375 TICK_ALLOC_UPD_PAP(words+1,0);
3377 /* now build o = FUN(seq,ap) */
3378 o = (StgClosure *)allocate(sizeofW(StgClosure)+1);
3379 TICK_ALLOC_SE_THK(1,0);
3380 SET_HDR(o,&stg_seq_info,su->header.prof.ccs /* ToDo */);
3381 o->payload[0] = (StgClosure *)ap;
3384 fprintf(stderr, "scheduler: Built ");
3385 printObj((StgClosure *)o);
3388 /* pop the old handler and put o on the stack */
3390 sp += sizeofW(StgSeqFrame) - 1;
3396 /* We've stripped the entire stack, the thread is now dead. */
3397 sp += sizeofW(StgStopFrame) - 1;
3398 sp[0] = (W_)exception; /* save the exception */
3399 tso->what_next = ThreadKilled;
3400 tso->su = (StgUpdateFrame *)(sp+1);
3411 /* -----------------------------------------------------------------------------
3412 resurrectThreads is called after garbage collection on the list of
3413 threads found to be garbage. Each of these threads will be woken
3414 up and sent a signal: BlockedOnDeadMVar if the thread was blocked
3415 on an MVar, or NonTermination if the thread was blocked on a Black
3418 Locks: sched_mutex isn't held upon entry nor exit.
3419 -------------------------------------------------------------------------- */
3422 resurrectThreads( StgTSO *threads )
3426 for (tso = threads; tso != END_TSO_QUEUE; tso = next) {
3427 next = tso->global_link;
3428 tso->global_link = all_threads;
3430 IF_DEBUG(scheduler, sched_belch("resurrecting thread %d", tso->id));
3432 switch (tso->why_blocked) {
3434 case BlockedOnException:
3435 /* Called by GC - sched_mutex lock is currently held. */
3436 raiseAsync(tso,(StgClosure *)BlockedOnDeadMVar_closure);
3438 case BlockedOnBlackHole:
3439 raiseAsync(tso,(StgClosure *)NonTermination_closure);
3442 /* This might happen if the thread was blocked on a black hole
3443 * belonging to a thread that we've just woken up (raiseAsync
3444 * can wake up threads, remember...).
3448 barf("resurrectThreads: thread blocked in a strange way");
3453 /* -----------------------------------------------------------------------------
3454 * Blackhole detection: if we reach a deadlock, test whether any
3455 * threads are blocked on themselves. Any threads which are found to
3456 * be self-blocked get sent a NonTermination exception.
3458 * This is only done in a deadlock situation in order to avoid
3459 * performance overhead in the normal case.
3461 * Locks: sched_mutex is held upon entry and exit.
3462 * -------------------------------------------------------------------------- */
3465 detectBlackHoles( void )
3467 StgTSO *t = all_threads;
3468 StgUpdateFrame *frame;
3469 StgClosure *blocked_on;
3471 for (t = all_threads; t != END_TSO_QUEUE; t = t->global_link) {
3473 while (t->what_next == ThreadRelocated) {
3475 ASSERT(get_itbl(t)->type == TSO);
3478 if (t->why_blocked != BlockedOnBlackHole) {
3482 blocked_on = t->block_info.closure;
3484 for (frame = t->su; ; frame = frame->link) {
3485 switch (get_itbl(frame)->type) {
3488 if (frame->updatee == blocked_on) {
3489 /* We are blocking on one of our own computations, so
3490 * send this thread the NonTermination exception.
3493 sched_belch("thread %d is blocked on itself", t->id));
3494 raiseAsync(t, (StgClosure *)NonTermination_closure);
3515 //@node Debugging Routines, Index, Exception Handling Routines, Main scheduling code
3516 //@subsection Debugging Routines
3518 /* -----------------------------------------------------------------------------
3519 Debugging: why is a thread blocked
3520 -------------------------------------------------------------------------- */
3525 printThreadBlockage(StgTSO *tso)
3527 switch (tso->why_blocked) {
3529 fprintf(stderr,"is blocked on read from fd %d", tso->block_info.fd);
3531 case BlockedOnWrite:
3532 fprintf(stderr,"is blocked on write to fd %d", tso->block_info.fd);
3534 case BlockedOnDelay:
3535 fprintf(stderr,"is blocked until %d", tso->block_info.target);
3538 fprintf(stderr,"is blocked on an MVar");
3540 case BlockedOnException:
3541 fprintf(stderr,"is blocked on delivering an exception to thread %d",
3542 tso->block_info.tso->id);
3544 case BlockedOnBlackHole:
3545 fprintf(stderr,"is blocked on a black hole");
3548 fprintf(stderr,"is not blocked");
3552 fprintf(stderr,"is blocked on global address; local FM_BQ is %p (%s)",
3553 tso->block_info.closure, info_type(tso->block_info.closure));
3555 case BlockedOnGA_NoSend:
3556 fprintf(stderr,"is blocked on global address (no send); local FM_BQ is %p (%s)",
3557 tso->block_info.closure, info_type(tso->block_info.closure));
3560 #if defined(RTS_SUPPORTS_THREADS)
3561 case BlockedOnCCall:
3562 fprintf(stderr,"is blocked on an external call");
3566 barf("printThreadBlockage: strange tso->why_blocked: %d for TSO %d (%d)",
3567 tso->why_blocked, tso->id, tso);
3572 printThreadStatus(StgTSO *tso)
3574 switch (tso->what_next) {
3576 fprintf(stderr,"has been killed");
3578 case ThreadComplete:
3579 fprintf(stderr,"has completed");
3582 printThreadBlockage(tso);
3587 printAllThreads(void)
3593 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3594 ullong_format_string(TIME_ON_PROC(CurrentProc),
3595 time_string, rtsFalse/*no commas!*/);
3597 sched_belch("all threads at [%s]:", time_string);
3599 char time_string[TIME_STR_LEN], node_str[NODE_STR_LEN];
3600 ullong_format_string(CURRENT_TIME,
3601 time_string, rtsFalse/*no commas!*/);
3603 sched_belch("all threads at [%s]:", time_string);
3605 sched_belch("all threads:");
3608 for (t = all_threads; t != END_TSO_QUEUE; t = t->global_link) {
3609 fprintf(stderr, "\tthread %d @ %p ", t->id, (void *)t);
3610 label = lookupThreadLabel((StgWord)t);
3611 if (label) fprintf(stderr,"[\"%s\"] ",(char *)label);
3612 printThreadStatus(t);
3613 fprintf(stderr,"\n");
3618 Print a whole blocking queue attached to node (debugging only).
3623 print_bq (StgClosure *node)
3625 StgBlockingQueueElement *bqe;
3629 fprintf(stderr,"## BQ of closure %p (%s): ",
3630 node, info_type(node));
3632 /* should cover all closures that may have a blocking queue */
3633 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3634 get_itbl(node)->type == FETCH_ME_BQ ||
3635 get_itbl(node)->type == RBH ||
3636 get_itbl(node)->type == MVAR);
3638 ASSERT(node!=(StgClosure*)NULL); // sanity check
3640 print_bqe(((StgBlockingQueue*)node)->blocking_queue);
3644 Print a whole blocking queue starting with the element bqe.
3647 print_bqe (StgBlockingQueueElement *bqe)
3652 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3654 for (end = (bqe==END_BQ_QUEUE);
3655 !end; // iterate until bqe points to a CONSTR
3656 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE),
3657 bqe = end ? END_BQ_QUEUE : bqe->link) {
3658 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3659 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3660 /* types of closures that may appear in a blocking queue */
3661 ASSERT(get_itbl(bqe)->type == TSO ||
3662 get_itbl(bqe)->type == BLOCKED_FETCH ||
3663 get_itbl(bqe)->type == CONSTR);
3664 /* only BQs of an RBH end with an RBH_Save closure */
3665 //ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3667 switch (get_itbl(bqe)->type) {
3669 fprintf(stderr," TSO %u (%x),",
3670 ((StgTSO *)bqe)->id, ((StgTSO *)bqe));
3673 fprintf(stderr," BF (node=%p, ga=((%x, %d, %x)),",
3674 ((StgBlockedFetch *)bqe)->node,
3675 ((StgBlockedFetch *)bqe)->ga.payload.gc.gtid,
3676 ((StgBlockedFetch *)bqe)->ga.payload.gc.slot,
3677 ((StgBlockedFetch *)bqe)->ga.weight);
3680 fprintf(stderr," %s (IP %p),",
3681 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3682 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3683 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3684 "RBH_Save_?"), get_itbl(bqe));
3687 barf("Unexpected closure type %s in blocking queue", // of %p (%s)",
3688 info_type((StgClosure *)bqe)); // , node, info_type(node));
3692 fputc('\n', stderr);
3694 # elif defined(GRAN)
3696 print_bq (StgClosure *node)
3698 StgBlockingQueueElement *bqe;
3699 PEs node_loc, tso_loc;
3702 /* should cover all closures that may have a blocking queue */
3703 ASSERT(get_itbl(node)->type == BLACKHOLE_BQ ||
3704 get_itbl(node)->type == FETCH_ME_BQ ||
3705 get_itbl(node)->type == RBH);
3707 ASSERT(node!=(StgClosure*)NULL); // sanity check
3708 node_loc = where_is(node);
3710 fprintf(stderr,"## BQ of closure %p (%s) on [PE %d]: ",
3711 node, info_type(node), node_loc);
3714 NB: In a parallel setup a BQ of an RBH must end with an RBH_Save closure;
3716 for (bqe = ((StgBlockingQueue*)node)->blocking_queue, end = (bqe==END_BQ_QUEUE);
3717 !end; // iterate until bqe points to a CONSTR
3718 end = (get_itbl(bqe)->type == CONSTR) || (bqe->link==END_BQ_QUEUE), bqe = end ? END_BQ_QUEUE : bqe->link) {
3719 ASSERT(bqe != END_BQ_QUEUE); // sanity check
3720 ASSERT(bqe != (StgBlockingQueueElement *)NULL); // sanity check
3721 /* types of closures that may appear in a blocking queue */
3722 ASSERT(get_itbl(bqe)->type == TSO ||
3723 get_itbl(bqe)->type == CONSTR);
3724 /* only BQs of an RBH end with an RBH_Save closure */
3725 ASSERT(get_itbl(bqe)->type != CONSTR || get_itbl(node)->type == RBH);
3727 tso_loc = where_is((StgClosure *)bqe);
3728 switch (get_itbl(bqe)->type) {
3730 fprintf(stderr," TSO %d (%p) on [PE %d],",
3731 ((StgTSO *)bqe)->id, (StgTSO *)bqe, tso_loc);
3734 fprintf(stderr," %s (IP %p),",
3735 (get_itbl(bqe) == &stg_RBH_Save_0_info ? "RBH_Save_0" :
3736 get_itbl(bqe) == &stg_RBH_Save_1_info ? "RBH_Save_1" :
3737 get_itbl(bqe) == &stg_RBH_Save_2_info ? "RBH_Save_2" :
3738 "RBH_Save_?"), get_itbl(bqe));
3741 barf("Unexpected closure type %s in blocking queue of %p (%s)",
3742 info_type((StgClosure *)bqe), node, info_type(node));
3746 fputc('\n', stderr);
3750 Nice and easy: only TSOs on the blocking queue
3753 print_bq (StgClosure *node)
3757 ASSERT(node!=(StgClosure*)NULL); // sanity check
3758 for (tso = ((StgBlockingQueue*)node)->blocking_queue;
3759 tso != END_TSO_QUEUE;
3761 ASSERT(tso!=NULL && tso!=END_TSO_QUEUE); // sanity check
3762 ASSERT(get_itbl(tso)->type == TSO); // guess what, sanity check
3763 fprintf(stderr," TSO %d (%p),", tso->id, tso);
3765 fputc('\n', stderr);
3776 for (i=0, tso=run_queue_hd;
3777 tso != END_TSO_QUEUE;
3786 sched_belch(char *s, ...)
3791 fprintf(stderr, "scheduler (task %ld): ", osThreadId());
3793 fprintf(stderr, "== ");
3795 fprintf(stderr, "scheduler: ");
3797 vfprintf(stderr, s, ap);
3798 fprintf(stderr, "\n");
3805 //@node Index, , Debugging Routines, Main scheduling code
3809 //* StgMainThread:: @cindex\s-+StgMainThread
3810 //* awaken_blocked_queue:: @cindex\s-+awaken_blocked_queue
3811 //* blocked_queue_hd:: @cindex\s-+blocked_queue_hd
3812 //* blocked_queue_tl:: @cindex\s-+blocked_queue_tl
3813 //* context_switch:: @cindex\s-+context_switch
3814 //* createThread:: @cindex\s-+createThread
3815 //* gc_pending_cond:: @cindex\s-+gc_pending_cond
3816 //* initScheduler:: @cindex\s-+initScheduler
3817 //* interrupted:: @cindex\s-+interrupted
3818 //* next_thread_id:: @cindex\s-+next_thread_id
3819 //* print_bq:: @cindex\s-+print_bq
3820 //* run_queue_hd:: @cindex\s-+run_queue_hd
3821 //* run_queue_tl:: @cindex\s-+run_queue_tl
3822 //* sched_mutex:: @cindex\s-+sched_mutex
3823 //* schedule:: @cindex\s-+schedule
3824 //* take_off_run_queue:: @cindex\s-+take_off_run_queue
3825 //* term_mutex:: @cindex\s-+term_mutex