Replace "tail -n +2" with "sed 1d", as Solaris doesn't understand the former
[ghc.git] / rts / Schedule.c
1 /* ---------------------------------------------------------------------------
2 *
3 * (c) The GHC Team, 1998-2006
4 *
5 * The scheduler and thread-related functionality
6 *
7 * --------------------------------------------------------------------------*/
8
9 #include "PosixSource.h"
10 #define KEEP_LOCKCLOSURE
11 #include "Rts.h"
12 #include "SchedAPI.h"
13 #include "RtsUtils.h"
14 #include "RtsFlags.h"
15 #include "OSThreads.h"
16 #include "Storage.h"
17 #include "StgRun.h"
18 #include "Hooks.h"
19 #include "Schedule.h"
20 #include "StgMiscClosures.h"
21 #include "Interpreter.h"
22 #include "Printer.h"
23 #include "RtsSignals.h"
24 #include "Sanity.h"
25 #include "Stats.h"
26 #include "STM.h"
27 #include "Timer.h"
28 #include "Prelude.h"
29 #include "ThreadLabels.h"
30 #include "LdvProfile.h"
31 #include "Updates.h"
32 #include "Proftimer.h"
33 #include "ProfHeap.h"
34 #if defined(GRAN) || defined(PARALLEL_HASKELL)
35 # include "GranSimRts.h"
36 # include "GranSim.h"
37 # include "ParallelRts.h"
38 # include "Parallel.h"
39 # include "ParallelDebug.h"
40 # include "FetchMe.h"
41 # include "HLC.h"
42 #endif
43 #include "Sparks.h"
44 #include "Capability.h"
45 #include "Task.h"
46 #include "AwaitEvent.h"
47 #if defined(mingw32_HOST_OS)
48 #include "win32/IOManager.h"
49 #endif
50 #include "Trace.h"
51 #include "RaiseAsync.h"
52 #include "Threads.h"
53 #include "ThrIOManager.h"
54
55 #ifdef HAVE_SYS_TYPES_H
56 #include <sys/types.h>
57 #endif
58 #ifdef HAVE_UNISTD_H
59 #include <unistd.h>
60 #endif
61
62 #include <string.h>
63 #include <stdlib.h>
64 #include <stdarg.h>
65
66 #ifdef HAVE_ERRNO_H
67 #include <errno.h>
68 #endif
69
70 // Turn off inlining when debugging - it obfuscates things
71 #ifdef DEBUG
72 # undef STATIC_INLINE
73 # define STATIC_INLINE static
74 #endif
75
76 /* -----------------------------------------------------------------------------
77 * Global variables
78 * -------------------------------------------------------------------------- */
79
80 #if defined(GRAN)
81
82 StgTSO* ActiveTSO = NULL; /* for assigning system costs; GranSim-Light only */
83 /* rtsTime TimeOfNextEvent, EndOfTimeSlice; now in GranSim.c */
84
85 /*
86 In GranSim we have a runnable and a blocked queue for each processor.
87 In order to minimise code changes new arrays run_queue_hds/tls
88 are created. run_queue_hd is then a short cut (macro) for
89 run_queue_hds[CurrentProc] (see GranSim.h).
90 -- HWL
91 */
92 StgTSO *run_queue_hds[MAX_PROC], *run_queue_tls[MAX_PROC];
93 StgTSO *blocked_queue_hds[MAX_PROC], *blocked_queue_tls[MAX_PROC];
94 StgTSO *ccalling_threadss[MAX_PROC];
95 /* We use the same global list of threads (all_threads) in GranSim as in
96 the std RTS (i.e. we are cheating). However, we don't use this list in
97 the GranSim specific code at the moment (so we are only potentially
98 cheating). */
99
100 #else /* !GRAN */
101
102 #if !defined(THREADED_RTS)
103 // Blocked/sleeping thrads
104 StgTSO *blocked_queue_hd = NULL;
105 StgTSO *blocked_queue_tl = NULL;
106 StgTSO *sleeping_queue = NULL; // perhaps replace with a hash table?
107 #endif
108
109 /* Threads blocked on blackholes.
110 * LOCK: sched_mutex+capability, or all capabilities
111 */
112 StgTSO *blackhole_queue = NULL;
113 #endif
114
115 /* The blackhole_queue should be checked for threads to wake up. See
116 * Schedule.h for more thorough comment.
117 * LOCK: none (doesn't matter if we miss an update)
118 */
119 rtsBool blackholes_need_checking = rtsFalse;
120
121 /* Linked list of all threads.
122 * Used for detecting garbage collected threads.
123 * LOCK: sched_mutex+capability, or all capabilities
124 */
125 StgTSO *all_threads = NULL;
126
127 /* flag set by signal handler to precipitate a context switch
128 * LOCK: none (just an advisory flag)
129 */
130 int context_switch = 0;
131
132 /* flag that tracks whether we have done any execution in this time slice.
133 * LOCK: currently none, perhaps we should lock (but needs to be
134 * updated in the fast path of the scheduler).
135 */
136 nat recent_activity = ACTIVITY_YES;
137
138 /* if this flag is set as well, give up execution
139 * LOCK: none (changes once, from false->true)
140 */
141 rtsBool sched_state = SCHED_RUNNING;
142
143 #if defined(GRAN)
144 StgTSO *CurrentTSO;
145 #endif
146
147 /* This is used in `TSO.h' and gcc 2.96 insists that this variable actually
148 * exists - earlier gccs apparently didn't.
149 * -= chak
150 */
151 StgTSO dummy_tso;
152
153 /*
154 * Set to TRUE when entering a shutdown state (via shutdownHaskellAndExit()) --
155 * in an MT setting, needed to signal that a worker thread shouldn't hang around
156 * in the scheduler when it is out of work.
157 */
158 rtsBool shutting_down_scheduler = rtsFalse;
159
160 /*
161 * This mutex protects most of the global scheduler data in
162 * the THREADED_RTS runtime.
163 */
164 #if defined(THREADED_RTS)
165 Mutex sched_mutex;
166 #endif
167
168 #if defined(PARALLEL_HASKELL)
169 StgTSO *LastTSO;
170 rtsTime TimeOfLastYield;
171 rtsBool emitSchedule = rtsTrue;
172 #endif
173
174 #if !defined(mingw32_HOST_OS)
175 #define FORKPROCESS_PRIMOP_SUPPORTED
176 #endif
177
178 /* -----------------------------------------------------------------------------
179 * static function prototypes
180 * -------------------------------------------------------------------------- */
181
182 static Capability *schedule (Capability *initialCapability, Task *task);
183
184 //
185 // These function all encapsulate parts of the scheduler loop, and are
186 // abstracted only to make the structure and control flow of the
187 // scheduler clearer.
188 //
189 static void schedulePreLoop (void);
190 #if defined(THREADED_RTS)
191 static void schedulePushWork(Capability *cap, Task *task);
192 #endif
193 static void scheduleStartSignalHandlers (Capability *cap);
194 static void scheduleCheckBlockedThreads (Capability *cap);
195 static void scheduleCheckWakeupThreads(Capability *cap USED_IF_NOT_THREADS);
196 static void scheduleCheckBlackHoles (Capability *cap);
197 static void scheduleDetectDeadlock (Capability *cap, Task *task);
198 #if defined(GRAN)
199 static StgTSO *scheduleProcessEvent(rtsEvent *event);
200 #endif
201 #if defined(PARALLEL_HASKELL)
202 static StgTSO *scheduleSendPendingMessages(void);
203 static void scheduleActivateSpark(void);
204 static rtsBool scheduleGetRemoteWork(rtsBool *receivedFinish);
205 #endif
206 #if defined(PAR) || defined(GRAN)
207 static void scheduleGranParReport(void);
208 #endif
209 static void schedulePostRunThread(void);
210 static rtsBool scheduleHandleHeapOverflow( Capability *cap, StgTSO *t );
211 static void scheduleHandleStackOverflow( Capability *cap, Task *task,
212 StgTSO *t);
213 static rtsBool scheduleHandleYield( Capability *cap, StgTSO *t,
214 nat prev_what_next );
215 static void scheduleHandleThreadBlocked( StgTSO *t );
216 static rtsBool scheduleHandleThreadFinished( Capability *cap, Task *task,
217 StgTSO *t );
218 static rtsBool scheduleNeedHeapProfile(rtsBool ready_to_gc);
219 static Capability *scheduleDoGC(Capability *cap, Task *task,
220 rtsBool force_major);
221
222 static rtsBool checkBlackHoles(Capability *cap);
223
224 static StgTSO *threadStackOverflow(Capability *cap, StgTSO *tso);
225
226 static void deleteThread (Capability *cap, StgTSO *tso);
227 static void deleteAllThreads (Capability *cap);
228
229 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
230 static void deleteThread_(Capability *cap, StgTSO *tso);
231 #endif
232
233 #if defined(PARALLEL_HASKELL)
234 StgTSO * createSparkThread(rtsSpark spark);
235 StgTSO * activateSpark (rtsSpark spark);
236 #endif
237
238 #ifdef DEBUG
239 static char *whatNext_strs[] = {
240 "(unknown)",
241 "ThreadRunGHC",
242 "ThreadInterpret",
243 "ThreadKilled",
244 "ThreadRelocated",
245 "ThreadComplete"
246 };
247 #endif
248
249 /* -----------------------------------------------------------------------------
250 * Putting a thread on the run queue: different scheduling policies
251 * -------------------------------------------------------------------------- */
252
253 STATIC_INLINE void
254 addToRunQueue( Capability *cap, StgTSO *t )
255 {
256 #if defined(PARALLEL_HASKELL)
257 if (RtsFlags.ParFlags.doFairScheduling) {
258 // this does round-robin scheduling; good for concurrency
259 appendToRunQueue(cap,t);
260 } else {
261 // this does unfair scheduling; good for parallelism
262 pushOnRunQueue(cap,t);
263 }
264 #else
265 // this does round-robin scheduling; good for concurrency
266 appendToRunQueue(cap,t);
267 #endif
268 }
269
270 /* ---------------------------------------------------------------------------
271 Main scheduling loop.
272
273 We use round-robin scheduling, each thread returning to the
274 scheduler loop when one of these conditions is detected:
275
276 * out of heap space
277 * timer expires (thread yields)
278 * thread blocks
279 * thread ends
280 * stack overflow
281
282 GRAN version:
283 In a GranSim setup this loop iterates over the global event queue.
284 This revolves around the global event queue, which determines what
285 to do next. Therefore, it's more complicated than either the
286 concurrent or the parallel (GUM) setup.
287
288 GUM version:
289 GUM iterates over incoming messages.
290 It starts with nothing to do (thus CurrentTSO == END_TSO_QUEUE),
291 and sends out a fish whenever it has nothing to do; in-between
292 doing the actual reductions (shared code below) it processes the
293 incoming messages and deals with delayed operations
294 (see PendingFetches).
295 This is not the ugliest code you could imagine, but it's bloody close.
296
297 ------------------------------------------------------------------------ */
298
299 static Capability *
300 schedule (Capability *initialCapability, Task *task)
301 {
302 StgTSO *t;
303 Capability *cap;
304 StgThreadReturnCode ret;
305 #if defined(GRAN)
306 rtsEvent *event;
307 #elif defined(PARALLEL_HASKELL)
308 StgTSO *tso;
309 GlobalTaskId pe;
310 rtsBool receivedFinish = rtsFalse;
311 # if defined(DEBUG)
312 nat tp_size, sp_size; // stats only
313 # endif
314 #endif
315 nat prev_what_next;
316 rtsBool ready_to_gc;
317 #if defined(THREADED_RTS)
318 rtsBool first = rtsTrue;
319 #endif
320
321 cap = initialCapability;
322
323 // Pre-condition: this task owns initialCapability.
324 // The sched_mutex is *NOT* held
325 // NB. on return, we still hold a capability.
326
327 debugTrace (DEBUG_sched,
328 "### NEW SCHEDULER LOOP (task: %p, cap: %p)",
329 task, initialCapability);
330
331 schedulePreLoop();
332
333 // -----------------------------------------------------------
334 // Scheduler loop starts here:
335
336 #if defined(PARALLEL_HASKELL)
337 #define TERMINATION_CONDITION (!receivedFinish)
338 #elif defined(GRAN)
339 #define TERMINATION_CONDITION ((event = get_next_event()) != (rtsEvent*)NULL)
340 #else
341 #define TERMINATION_CONDITION rtsTrue
342 #endif
343
344 while (TERMINATION_CONDITION) {
345
346 #if defined(GRAN)
347 /* Choose the processor with the next event */
348 CurrentProc = event->proc;
349 CurrentTSO = event->tso;
350 #endif
351
352 #if defined(THREADED_RTS)
353 if (first) {
354 // don't yield the first time, we want a chance to run this
355 // thread for a bit, even if there are others banging at the
356 // door.
357 first = rtsFalse;
358 ASSERT_FULL_CAPABILITY_INVARIANTS(cap,task);
359 } else {
360 // Yield the capability to higher-priority tasks if necessary.
361 yieldCapability(&cap, task);
362 }
363 #endif
364
365 #if defined(THREADED_RTS)
366 schedulePushWork(cap,task);
367 #endif
368
369 // Check whether we have re-entered the RTS from Haskell without
370 // going via suspendThread()/resumeThread (i.e. a 'safe' foreign
371 // call).
372 if (cap->in_haskell) {
373 errorBelch("schedule: re-entered unsafely.\n"
374 " Perhaps a 'foreign import unsafe' should be 'safe'?");
375 stg_exit(EXIT_FAILURE);
376 }
377
378 // The interruption / shutdown sequence.
379 //
380 // In order to cleanly shut down the runtime, we want to:
381 // * make sure that all main threads return to their callers
382 // with the state 'Interrupted'.
383 // * clean up all OS threads assocated with the runtime
384 // * free all memory etc.
385 //
386 // So the sequence for ^C goes like this:
387 //
388 // * ^C handler sets sched_state := SCHED_INTERRUPTING and
389 // arranges for some Capability to wake up
390 //
391 // * all threads in the system are halted, and the zombies are
392 // placed on the run queue for cleaning up. We acquire all
393 // the capabilities in order to delete the threads, this is
394 // done by scheduleDoGC() for convenience (because GC already
395 // needs to acquire all the capabilities). We can't kill
396 // threads involved in foreign calls.
397 //
398 // * somebody calls shutdownHaskell(), which calls exitScheduler()
399 //
400 // * sched_state := SCHED_SHUTTING_DOWN
401 //
402 // * all workers exit when the run queue on their capability
403 // drains. All main threads will also exit when their TSO
404 // reaches the head of the run queue and they can return.
405 //
406 // * eventually all Capabilities will shut down, and the RTS can
407 // exit.
408 //
409 // * We might be left with threads blocked in foreign calls,
410 // we should really attempt to kill these somehow (TODO);
411
412 switch (sched_state) {
413 case SCHED_RUNNING:
414 break;
415 case SCHED_INTERRUPTING:
416 debugTrace(DEBUG_sched, "SCHED_INTERRUPTING");
417 #if defined(THREADED_RTS)
418 discardSparksCap(cap);
419 #endif
420 /* scheduleDoGC() deletes all the threads */
421 cap = scheduleDoGC(cap,task,rtsFalse);
422 break;
423 case SCHED_SHUTTING_DOWN:
424 debugTrace(DEBUG_sched, "SCHED_SHUTTING_DOWN");
425 // If we are a worker, just exit. If we're a bound thread
426 // then we will exit below when we've removed our TSO from
427 // the run queue.
428 if (task->tso == NULL && emptyRunQueue(cap)) {
429 return cap;
430 }
431 break;
432 default:
433 barf("sched_state: %d", sched_state);
434 }
435
436 #if defined(THREADED_RTS)
437 // If the run queue is empty, take a spark and turn it into a thread.
438 {
439 if (emptyRunQueue(cap)) {
440 StgClosure *spark;
441 spark = findSpark(cap);
442 if (spark != NULL) {
443 debugTrace(DEBUG_sched,
444 "turning spark of closure %p into a thread",
445 (StgClosure *)spark);
446 createSparkThread(cap,spark);
447 }
448 }
449 }
450 #endif // THREADED_RTS
451
452 scheduleStartSignalHandlers(cap);
453
454 // Only check the black holes here if we've nothing else to do.
455 // During normal execution, the black hole list only gets checked
456 // at GC time, to avoid repeatedly traversing this possibly long
457 // list each time around the scheduler.
458 if (emptyRunQueue(cap)) { scheduleCheckBlackHoles(cap); }
459
460 scheduleCheckWakeupThreads(cap);
461
462 scheduleCheckBlockedThreads(cap);
463
464 scheduleDetectDeadlock(cap,task);
465 #if defined(THREADED_RTS)
466 cap = task->cap; // reload cap, it might have changed
467 #endif
468
469 // Normally, the only way we can get here with no threads to
470 // run is if a keyboard interrupt received during
471 // scheduleCheckBlockedThreads() or scheduleDetectDeadlock().
472 // Additionally, it is not fatal for the
473 // threaded RTS to reach here with no threads to run.
474 //
475 // win32: might be here due to awaitEvent() being abandoned
476 // as a result of a console event having been delivered.
477 if ( emptyRunQueue(cap) ) {
478 #if !defined(THREADED_RTS) && !defined(mingw32_HOST_OS)
479 ASSERT(sched_state >= SCHED_INTERRUPTING);
480 #endif
481 continue; // nothing to do
482 }
483
484 #if defined(PARALLEL_HASKELL)
485 scheduleSendPendingMessages();
486 if (emptyRunQueue(cap) && scheduleActivateSpark())
487 continue;
488
489 #if defined(SPARKS)
490 ASSERT(next_fish_to_send_at==0); // i.e. no delayed fishes left!
491 #endif
492
493 /* If we still have no work we need to send a FISH to get a spark
494 from another PE */
495 if (emptyRunQueue(cap)) {
496 if (!scheduleGetRemoteWork(&receivedFinish)) continue;
497 ASSERT(rtsFalse); // should not happen at the moment
498 }
499 // from here: non-empty run queue.
500 // TODO: merge above case with this, only one call processMessages() !
501 if (PacketsWaiting()) { /* process incoming messages, if
502 any pending... only in else
503 because getRemoteWork waits for
504 messages as well */
505 receivedFinish = processMessages();
506 }
507 #endif
508
509 #if defined(GRAN)
510 scheduleProcessEvent(event);
511 #endif
512
513 //
514 // Get a thread to run
515 //
516 t = popRunQueue(cap);
517
518 #if defined(GRAN) || defined(PAR)
519 scheduleGranParReport(); // some kind of debuging output
520 #else
521 // Sanity check the thread we're about to run. This can be
522 // expensive if there is lots of thread switching going on...
523 IF_DEBUG(sanity,checkTSO(t));
524 #endif
525
526 #if defined(THREADED_RTS)
527 // Check whether we can run this thread in the current task.
528 // If not, we have to pass our capability to the right task.
529 {
530 Task *bound = t->bound;
531
532 if (bound) {
533 if (bound == task) {
534 debugTrace(DEBUG_sched,
535 "### Running thread %lu in bound thread", (unsigned long)t->id);
536 // yes, the Haskell thread is bound to the current native thread
537 } else {
538 debugTrace(DEBUG_sched,
539 "### thread %lu bound to another OS thread", (unsigned long)t->id);
540 // no, bound to a different Haskell thread: pass to that thread
541 pushOnRunQueue(cap,t);
542 continue;
543 }
544 } else {
545 // The thread we want to run is unbound.
546 if (task->tso) {
547 debugTrace(DEBUG_sched,
548 "### this OS thread cannot run thread %lu", (unsigned long)t->id);
549 // no, the current native thread is bound to a different
550 // Haskell thread, so pass it to any worker thread
551 pushOnRunQueue(cap,t);
552 continue;
553 }
554 }
555 }
556 #endif
557
558 cap->r.rCurrentTSO = t;
559
560 /* context switches are initiated by the timer signal, unless
561 * the user specified "context switch as often as possible", with
562 * +RTS -C0
563 */
564 if (RtsFlags.ConcFlags.ctxtSwitchTicks == 0
565 && !emptyThreadQueues(cap)) {
566 context_switch = 1;
567 }
568
569 run_thread:
570
571 debugTrace(DEBUG_sched, "-->> running thread %ld %s ...",
572 (long)t->id, whatNext_strs[t->what_next]);
573
574 startHeapProfTimer();
575
576 // Check for exceptions blocked on this thread
577 maybePerformBlockedException (cap, t);
578
579 // ----------------------------------------------------------------------
580 // Run the current thread
581
582 ASSERT_FULL_CAPABILITY_INVARIANTS(cap,task);
583 ASSERT(t->cap == cap);
584
585 prev_what_next = t->what_next;
586
587 errno = t->saved_errno;
588 #if mingw32_HOST_OS
589 SetLastError(t->saved_winerror);
590 #endif
591
592 cap->in_haskell = rtsTrue;
593
594 dirtyTSO(t);
595
596 #if defined(THREADED_RTS)
597 if (recent_activity == ACTIVITY_DONE_GC) {
598 // ACTIVITY_DONE_GC means we turned off the timer signal to
599 // conserve power (see #1623). Re-enable it here.
600 nat prev;
601 prev = xchg(&recent_activity, ACTIVITY_YES);
602 if (prev == ACTIVITY_DONE_GC) {
603 startTimer();
604 }
605 } else {
606 recent_activity = ACTIVITY_YES;
607 }
608 #endif
609
610 switch (prev_what_next) {
611
612 case ThreadKilled:
613 case ThreadComplete:
614 /* Thread already finished, return to scheduler. */
615 ret = ThreadFinished;
616 break;
617
618 case ThreadRunGHC:
619 {
620 StgRegTable *r;
621 r = StgRun((StgFunPtr) stg_returnToStackTop, &cap->r);
622 cap = regTableToCapability(r);
623 ret = r->rRet;
624 break;
625 }
626
627 case ThreadInterpret:
628 cap = interpretBCO(cap);
629 ret = cap->r.rRet;
630 break;
631
632 default:
633 barf("schedule: invalid what_next field");
634 }
635
636 cap->in_haskell = rtsFalse;
637
638 // The TSO might have moved, eg. if it re-entered the RTS and a GC
639 // happened. So find the new location:
640 t = cap->r.rCurrentTSO;
641
642 // We have run some Haskell code: there might be blackhole-blocked
643 // threads to wake up now.
644 // Lock-free test here should be ok, we're just setting a flag.
645 if ( blackhole_queue != END_TSO_QUEUE ) {
646 blackholes_need_checking = rtsTrue;
647 }
648
649 // And save the current errno in this thread.
650 // XXX: possibly bogus for SMP because this thread might already
651 // be running again, see code below.
652 t->saved_errno = errno;
653 #if mingw32_HOST_OS
654 // Similarly for Windows error code
655 t->saved_winerror = GetLastError();
656 #endif
657
658 #if defined(THREADED_RTS)
659 // If ret is ThreadBlocked, and this Task is bound to the TSO that
660 // blocked, we are in limbo - the TSO is now owned by whatever it
661 // is blocked on, and may in fact already have been woken up,
662 // perhaps even on a different Capability. It may be the case
663 // that task->cap != cap. We better yield this Capability
664 // immediately and return to normaility.
665 if (ret == ThreadBlocked) {
666 debugTrace(DEBUG_sched,
667 "--<< thread %lu (%s) stopped: blocked",
668 (unsigned long)t->id, whatNext_strs[t->what_next]);
669 continue;
670 }
671 #endif
672
673 ASSERT_FULL_CAPABILITY_INVARIANTS(cap,task);
674 ASSERT(t->cap == cap);
675
676 // ----------------------------------------------------------------------
677
678 // Costs for the scheduler are assigned to CCS_SYSTEM
679 stopHeapProfTimer();
680 #if defined(PROFILING)
681 CCCS = CCS_SYSTEM;
682 #endif
683
684 schedulePostRunThread();
685
686 ready_to_gc = rtsFalse;
687
688 switch (ret) {
689 case HeapOverflow:
690 ready_to_gc = scheduleHandleHeapOverflow(cap,t);
691 break;
692
693 case StackOverflow:
694 scheduleHandleStackOverflow(cap,task,t);
695 break;
696
697 case ThreadYielding:
698 if (scheduleHandleYield(cap, t, prev_what_next)) {
699 // shortcut for switching between compiler/interpreter:
700 goto run_thread;
701 }
702 break;
703
704 case ThreadBlocked:
705 scheduleHandleThreadBlocked(t);
706 break;
707
708 case ThreadFinished:
709 if (scheduleHandleThreadFinished(cap, task, t)) return cap;
710 ASSERT_FULL_CAPABILITY_INVARIANTS(cap,task);
711 break;
712
713 default:
714 barf("schedule: invalid thread return code %d", (int)ret);
715 }
716
717 if (ready_to_gc || scheduleNeedHeapProfile(ready_to_gc)) {
718 cap = scheduleDoGC(cap,task,rtsFalse);
719 }
720 } /* end of while() */
721
722 debugTrace(PAR_DEBUG_verbose,
723 "== Leaving schedule() after having received Finish");
724 }
725
726 /* ----------------------------------------------------------------------------
727 * Setting up the scheduler loop
728 * ------------------------------------------------------------------------- */
729
730 static void
731 schedulePreLoop(void)
732 {
733 #if defined(GRAN)
734 /* set up first event to get things going */
735 /* ToDo: assign costs for system setup and init MainTSO ! */
736 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
737 ContinueThread,
738 CurrentTSO, (StgClosure*)NULL, (rtsSpark*)NULL);
739
740 debugTrace (DEBUG_gran,
741 "GRAN: Init CurrentTSO (in schedule) = %p",
742 CurrentTSO);
743 IF_DEBUG(gran, G_TSO(CurrentTSO, 5));
744
745 if (RtsFlags.GranFlags.Light) {
746 /* Save current time; GranSim Light only */
747 CurrentTSO->gran.clock = CurrentTime[CurrentProc];
748 }
749 #endif
750 }
751
752 /* -----------------------------------------------------------------------------
753 * schedulePushWork()
754 *
755 * Push work to other Capabilities if we have some.
756 * -------------------------------------------------------------------------- */
757
758 #if defined(THREADED_RTS)
759 static void
760 schedulePushWork(Capability *cap USED_IF_THREADS,
761 Task *task USED_IF_THREADS)
762 {
763 Capability *free_caps[n_capabilities], *cap0;
764 nat i, n_free_caps;
765
766 // migration can be turned off with +RTS -qg
767 if (!RtsFlags.ParFlags.migrate) return;
768
769 // Check whether we have more threads on our run queue, or sparks
770 // in our pool, that we could hand to another Capability.
771 if ((emptyRunQueue(cap) || cap->run_queue_hd->link == END_TSO_QUEUE)
772 && sparkPoolSizeCap(cap) < 2) {
773 return;
774 }
775
776 // First grab as many free Capabilities as we can.
777 for (i=0, n_free_caps=0; i < n_capabilities; i++) {
778 cap0 = &capabilities[i];
779 if (cap != cap0 && tryGrabCapability(cap0,task)) {
780 if (!emptyRunQueue(cap0) || cap->returning_tasks_hd != NULL) {
781 // it already has some work, we just grabbed it at
782 // the wrong moment. Or maybe it's deadlocked!
783 releaseCapability(cap0);
784 } else {
785 free_caps[n_free_caps++] = cap0;
786 }
787 }
788 }
789
790 // we now have n_free_caps free capabilities stashed in
791 // free_caps[]. Share our run queue equally with them. This is
792 // probably the simplest thing we could do; improvements we might
793 // want to do include:
794 //
795 // - giving high priority to moving relatively new threads, on
796 // the gournds that they haven't had time to build up a
797 // working set in the cache on this CPU/Capability.
798 //
799 // - giving low priority to moving long-lived threads
800
801 if (n_free_caps > 0) {
802 StgTSO *prev, *t, *next;
803 rtsBool pushed_to_all;
804
805 debugTrace(DEBUG_sched, "excess threads on run queue and %d free capabilities, sharing...", n_free_caps);
806
807 i = 0;
808 pushed_to_all = rtsFalse;
809
810 if (cap->run_queue_hd != END_TSO_QUEUE) {
811 prev = cap->run_queue_hd;
812 t = prev->link;
813 prev->link = END_TSO_QUEUE;
814 for (; t != END_TSO_QUEUE; t = next) {
815 next = t->link;
816 t->link = END_TSO_QUEUE;
817 if (t->what_next == ThreadRelocated
818 || t->bound == task // don't move my bound thread
819 || tsoLocked(t)) { // don't move a locked thread
820 prev->link = t;
821 prev = t;
822 } else if (i == n_free_caps) {
823 pushed_to_all = rtsTrue;
824 i = 0;
825 // keep one for us
826 prev->link = t;
827 prev = t;
828 } else {
829 debugTrace(DEBUG_sched, "pushing thread %lu to capability %d", (unsigned long)t->id, free_caps[i]->no);
830 appendToRunQueue(free_caps[i],t);
831 if (t->bound) { t->bound->cap = free_caps[i]; }
832 t->cap = free_caps[i];
833 i++;
834 }
835 }
836 cap->run_queue_tl = prev;
837 }
838
839 // If there are some free capabilities that we didn't push any
840 // threads to, then try to push a spark to each one.
841 if (!pushed_to_all) {
842 StgClosure *spark;
843 // i is the next free capability to push to
844 for (; i < n_free_caps; i++) {
845 if (emptySparkPoolCap(free_caps[i])) {
846 spark = findSpark(cap);
847 if (spark != NULL) {
848 debugTrace(DEBUG_sched, "pushing spark %p to capability %d", spark, free_caps[i]->no);
849 newSpark(&(free_caps[i]->r), spark);
850 }
851 }
852 }
853 }
854
855 // release the capabilities
856 for (i = 0; i < n_free_caps; i++) {
857 task->cap = free_caps[i];
858 releaseCapability(free_caps[i]);
859 }
860 }
861 task->cap = cap; // reset to point to our Capability.
862 }
863 #endif
864
865 /* ----------------------------------------------------------------------------
866 * Start any pending signal handlers
867 * ------------------------------------------------------------------------- */
868
869 #if defined(RTS_USER_SIGNALS) && !defined(THREADED_RTS)
870 static void
871 scheduleStartSignalHandlers(Capability *cap)
872 {
873 if (RtsFlags.MiscFlags.install_signal_handlers && signals_pending()) {
874 // safe outside the lock
875 startSignalHandlers(cap);
876 }
877 }
878 #else
879 static void
880 scheduleStartSignalHandlers(Capability *cap STG_UNUSED)
881 {
882 }
883 #endif
884
885 /* ----------------------------------------------------------------------------
886 * Check for blocked threads that can be woken up.
887 * ------------------------------------------------------------------------- */
888
889 static void
890 scheduleCheckBlockedThreads(Capability *cap USED_IF_NOT_THREADS)
891 {
892 #if !defined(THREADED_RTS)
893 //
894 // Check whether any waiting threads need to be woken up. If the
895 // run queue is empty, and there are no other tasks running, we
896 // can wait indefinitely for something to happen.
897 //
898 if ( !emptyQueue(blocked_queue_hd) || !emptyQueue(sleeping_queue) )
899 {
900 awaitEvent( emptyRunQueue(cap) && !blackholes_need_checking );
901 }
902 #endif
903 }
904
905
906 /* ----------------------------------------------------------------------------
907 * Check for threads woken up by other Capabilities
908 * ------------------------------------------------------------------------- */
909
910 static void
911 scheduleCheckWakeupThreads(Capability *cap USED_IF_THREADS)
912 {
913 #if defined(THREADED_RTS)
914 // Any threads that were woken up by other Capabilities get
915 // appended to our run queue.
916 if (!emptyWakeupQueue(cap)) {
917 ACQUIRE_LOCK(&cap->lock);
918 if (emptyRunQueue(cap)) {
919 cap->run_queue_hd = cap->wakeup_queue_hd;
920 cap->run_queue_tl = cap->wakeup_queue_tl;
921 } else {
922 cap->run_queue_tl->link = cap->wakeup_queue_hd;
923 cap->run_queue_tl = cap->wakeup_queue_tl;
924 }
925 cap->wakeup_queue_hd = cap->wakeup_queue_tl = END_TSO_QUEUE;
926 RELEASE_LOCK(&cap->lock);
927 }
928 #endif
929 }
930
931 /* ----------------------------------------------------------------------------
932 * Check for threads blocked on BLACKHOLEs that can be woken up
933 * ------------------------------------------------------------------------- */
934 static void
935 scheduleCheckBlackHoles (Capability *cap)
936 {
937 if ( blackholes_need_checking ) // check without the lock first
938 {
939 ACQUIRE_LOCK(&sched_mutex);
940 if ( blackholes_need_checking ) {
941 checkBlackHoles(cap);
942 blackholes_need_checking = rtsFalse;
943 }
944 RELEASE_LOCK(&sched_mutex);
945 }
946 }
947
948 /* ----------------------------------------------------------------------------
949 * Detect deadlock conditions and attempt to resolve them.
950 * ------------------------------------------------------------------------- */
951
952 static void
953 scheduleDetectDeadlock (Capability *cap, Task *task)
954 {
955
956 #if defined(PARALLEL_HASKELL)
957 // ToDo: add deadlock detection in GUM (similar to THREADED_RTS) -- HWL
958 return;
959 #endif
960
961 /*
962 * Detect deadlock: when we have no threads to run, there are no
963 * threads blocked, waiting for I/O, or sleeping, and all the
964 * other tasks are waiting for work, we must have a deadlock of
965 * some description.
966 */
967 if ( emptyThreadQueues(cap) )
968 {
969 #if defined(THREADED_RTS)
970 /*
971 * In the threaded RTS, we only check for deadlock if there
972 * has been no activity in a complete timeslice. This means
973 * we won't eagerly start a full GC just because we don't have
974 * any threads to run currently.
975 */
976 if (recent_activity != ACTIVITY_INACTIVE) return;
977 #endif
978
979 debugTrace(DEBUG_sched, "deadlocked, forcing major GC...");
980
981 // Garbage collection can release some new threads due to
982 // either (a) finalizers or (b) threads resurrected because
983 // they are unreachable and will therefore be sent an
984 // exception. Any threads thus released will be immediately
985 // runnable.
986 cap = scheduleDoGC (cap, task, rtsTrue/*force major GC*/);
987
988 recent_activity = ACTIVITY_DONE_GC;
989 // disable timer signals (see #1623)
990 stopTimer();
991
992 if ( !emptyRunQueue(cap) ) return;
993
994 #if defined(RTS_USER_SIGNALS) && !defined(THREADED_RTS)
995 /* If we have user-installed signal handlers, then wait
996 * for signals to arrive rather then bombing out with a
997 * deadlock.
998 */
999 if ( RtsFlags.MiscFlags.install_signal_handlers && anyUserHandlers() ) {
1000 debugTrace(DEBUG_sched,
1001 "still deadlocked, waiting for signals...");
1002
1003 awaitUserSignals();
1004
1005 if (signals_pending()) {
1006 startSignalHandlers(cap);
1007 }
1008
1009 // either we have threads to run, or we were interrupted:
1010 ASSERT(!emptyRunQueue(cap) || sched_state >= SCHED_INTERRUPTING);
1011 }
1012 #endif
1013
1014 #if !defined(THREADED_RTS)
1015 /* Probably a real deadlock. Send the current main thread the
1016 * Deadlock exception.
1017 */
1018 if (task->tso) {
1019 switch (task->tso->why_blocked) {
1020 case BlockedOnSTM:
1021 case BlockedOnBlackHole:
1022 case BlockedOnException:
1023 case BlockedOnMVar:
1024 throwToSingleThreaded(cap, task->tso,
1025 (StgClosure *)NonTermination_closure);
1026 return;
1027 default:
1028 barf("deadlock: main thread blocked in a strange way");
1029 }
1030 }
1031 return;
1032 #endif
1033 }
1034 }
1035
1036 /* ----------------------------------------------------------------------------
1037 * Process an event (GRAN only)
1038 * ------------------------------------------------------------------------- */
1039
1040 #if defined(GRAN)
1041 static StgTSO *
1042 scheduleProcessEvent(rtsEvent *event)
1043 {
1044 StgTSO *t;
1045
1046 if (RtsFlags.GranFlags.Light)
1047 GranSimLight_enter_system(event, &ActiveTSO); // adjust ActiveTSO etc
1048
1049 /* adjust time based on time-stamp */
1050 if (event->time > CurrentTime[CurrentProc] &&
1051 event->evttype != ContinueThread)
1052 CurrentTime[CurrentProc] = event->time;
1053
1054 /* Deal with the idle PEs (may issue FindWork or MoveSpark events) */
1055 if (!RtsFlags.GranFlags.Light)
1056 handleIdlePEs();
1057
1058 IF_DEBUG(gran, debugBelch("GRAN: switch by event-type\n"));
1059
1060 /* main event dispatcher in GranSim */
1061 switch (event->evttype) {
1062 /* Should just be continuing execution */
1063 case ContinueThread:
1064 IF_DEBUG(gran, debugBelch("GRAN: doing ContinueThread\n"));
1065 /* ToDo: check assertion
1066 ASSERT(run_queue_hd != (StgTSO*)NULL &&
1067 run_queue_hd != END_TSO_QUEUE);
1068 */
1069 /* Ignore ContinueThreads for fetching threads (if synchr comm) */
1070 if (!RtsFlags.GranFlags.DoAsyncFetch &&
1071 procStatus[CurrentProc]==Fetching) {
1072 debugBelch("ghuH: Spurious ContinueThread while Fetching ignored; TSO %d (%p) [PE %d]\n",
1073 CurrentTSO->id, CurrentTSO, CurrentProc);
1074 goto next_thread;
1075 }
1076 /* Ignore ContinueThreads for completed threads */
1077 if (CurrentTSO->what_next == ThreadComplete) {
1078 debugBelch("ghuH: found a ContinueThread event for completed thread %d (%p) [PE %d] (ignoring ContinueThread)\n",
1079 CurrentTSO->id, CurrentTSO, CurrentProc);
1080 goto next_thread;
1081 }
1082 /* Ignore ContinueThreads for threads that are being migrated */
1083 if (PROCS(CurrentTSO)==Nowhere) {
1084 debugBelch("ghuH: trying to run the migrating TSO %d (%p) [PE %d] (ignoring ContinueThread)\n",
1085 CurrentTSO->id, CurrentTSO, CurrentProc);
1086 goto next_thread;
1087 }
1088 /* The thread should be at the beginning of the run queue */
1089 if (CurrentTSO!=run_queue_hds[CurrentProc]) {
1090 debugBelch("ghuH: TSO %d (%p) [PE %d] is not at the start of the run_queue when doing a ContinueThread\n",
1091 CurrentTSO->id, CurrentTSO, CurrentProc);
1092 break; // run the thread anyway
1093 }
1094 /*
1095 new_event(proc, proc, CurrentTime[proc],
1096 FindWork,
1097 (StgTSO*)NULL, (StgClosure*)NULL, (rtsSpark*)NULL);
1098 goto next_thread;
1099 */ /* Catches superfluous CONTINUEs -- should be unnecessary */
1100 break; // now actually run the thread; DaH Qu'vam yImuHbej
1101
1102 case FetchNode:
1103 do_the_fetchnode(event);
1104 goto next_thread; /* handle next event in event queue */
1105
1106 case GlobalBlock:
1107 do_the_globalblock(event);
1108 goto next_thread; /* handle next event in event queue */
1109
1110 case FetchReply:
1111 do_the_fetchreply(event);
1112 goto next_thread; /* handle next event in event queue */
1113
1114 case UnblockThread: /* Move from the blocked queue to the tail of */
1115 do_the_unblock(event);
1116 goto next_thread; /* handle next event in event queue */
1117
1118 case ResumeThread: /* Move from the blocked queue to the tail of */
1119 /* the runnable queue ( i.e. Qu' SImqa'lu') */
1120 event->tso->gran.blocktime +=
1121 CurrentTime[CurrentProc] - event->tso->gran.blockedat;
1122 do_the_startthread(event);
1123 goto next_thread; /* handle next event in event queue */
1124
1125 case StartThread:
1126 do_the_startthread(event);
1127 goto next_thread; /* handle next event in event queue */
1128
1129 case MoveThread:
1130 do_the_movethread(event);
1131 goto next_thread; /* handle next event in event queue */
1132
1133 case MoveSpark:
1134 do_the_movespark(event);
1135 goto next_thread; /* handle next event in event queue */
1136
1137 case FindWork:
1138 do_the_findwork(event);
1139 goto next_thread; /* handle next event in event queue */
1140
1141 default:
1142 barf("Illegal event type %u\n", event->evttype);
1143 } /* switch */
1144
1145 /* This point was scheduler_loop in the old RTS */
1146
1147 IF_DEBUG(gran, debugBelch("GRAN: after main switch\n"));
1148
1149 TimeOfLastEvent = CurrentTime[CurrentProc];
1150 TimeOfNextEvent = get_time_of_next_event();
1151 IgnoreEvents=(TimeOfNextEvent==0); // HWL HACK
1152 // CurrentTSO = ThreadQueueHd;
1153
1154 IF_DEBUG(gran, debugBelch("GRAN: time of next event is: %ld\n",
1155 TimeOfNextEvent));
1156
1157 if (RtsFlags.GranFlags.Light)
1158 GranSimLight_leave_system(event, &ActiveTSO);
1159
1160 EndOfTimeSlice = CurrentTime[CurrentProc]+RtsFlags.GranFlags.time_slice;
1161
1162 IF_DEBUG(gran,
1163 debugBelch("GRAN: end of time-slice is %#lx\n", EndOfTimeSlice));
1164
1165 /* in a GranSim setup the TSO stays on the run queue */
1166 t = CurrentTSO;
1167 /* Take a thread from the run queue. */
1168 POP_RUN_QUEUE(t); // take_off_run_queue(t);
1169
1170 IF_DEBUG(gran,
1171 debugBelch("GRAN: About to run current thread, which is\n");
1172 G_TSO(t,5));
1173
1174 context_switch = 0; // turned on via GranYield, checking events and time slice
1175
1176 IF_DEBUG(gran,
1177 DumpGranEvent(GR_SCHEDULE, t));
1178
1179 procStatus[CurrentProc] = Busy;
1180 }
1181 #endif // GRAN
1182
1183 /* ----------------------------------------------------------------------------
1184 * Send pending messages (PARALLEL_HASKELL only)
1185 * ------------------------------------------------------------------------- */
1186
1187 #if defined(PARALLEL_HASKELL)
1188 static StgTSO *
1189 scheduleSendPendingMessages(void)
1190 {
1191 StgSparkPool *pool;
1192 rtsSpark spark;
1193 StgTSO *t;
1194
1195 # if defined(PAR) // global Mem.Mgmt., omit for now
1196 if (PendingFetches != END_BF_QUEUE) {
1197 processFetches();
1198 }
1199 # endif
1200
1201 if (RtsFlags.ParFlags.BufferTime) {
1202 // if we use message buffering, we must send away all message
1203 // packets which have become too old...
1204 sendOldBuffers();
1205 }
1206 }
1207 #endif
1208
1209 /* ----------------------------------------------------------------------------
1210 * Activate spark threads (PARALLEL_HASKELL only)
1211 * ------------------------------------------------------------------------- */
1212
1213 #if defined(PARALLEL_HASKELL)
1214 static void
1215 scheduleActivateSpark(void)
1216 {
1217 #if defined(SPARKS)
1218 ASSERT(emptyRunQueue());
1219 /* We get here if the run queue is empty and want some work.
1220 We try to turn a spark into a thread, and add it to the run queue,
1221 from where it will be picked up in the next iteration of the scheduler
1222 loop.
1223 */
1224
1225 /* :-[ no local threads => look out for local sparks */
1226 /* the spark pool for the current PE */
1227 pool = &(cap.r.rSparks); // JB: cap = (old) MainCap
1228 if (advisory_thread_count < RtsFlags.ParFlags.maxThreads &&
1229 pool->hd < pool->tl) {
1230 /*
1231 * ToDo: add GC code check that we really have enough heap afterwards!!
1232 * Old comment:
1233 * If we're here (no runnable threads) and we have pending
1234 * sparks, we must have a space problem. Get enough space
1235 * to turn one of those pending sparks into a
1236 * thread...
1237 */
1238
1239 spark = findSpark(rtsFalse); /* get a spark */
1240 if (spark != (rtsSpark) NULL) {
1241 tso = createThreadFromSpark(spark); /* turn the spark into a thread */
1242 IF_PAR_DEBUG(fish, // schedule,
1243 debugBelch("==== schedule: Created TSO %d (%p); %d threads active\n",
1244 tso->id, tso, advisory_thread_count));
1245
1246 if (tso==END_TSO_QUEUE) { /* failed to activate spark->back to loop */
1247 IF_PAR_DEBUG(fish, // schedule,
1248 debugBelch("==^^ failed to create thread from spark @ %lx\n",
1249 spark));
1250 return rtsFalse; /* failed to generate a thread */
1251 } /* otherwise fall through & pick-up new tso */
1252 } else {
1253 IF_PAR_DEBUG(fish, // schedule,
1254 debugBelch("==^^ no local sparks (spark pool contains only NFs: %d)\n",
1255 spark_queue_len(pool)));
1256 return rtsFalse; /* failed to generate a thread */
1257 }
1258 return rtsTrue; /* success in generating a thread */
1259 } else { /* no more threads permitted or pool empty */
1260 return rtsFalse; /* failed to generateThread */
1261 }
1262 #else
1263 tso = NULL; // avoid compiler warning only
1264 return rtsFalse; /* dummy in non-PAR setup */
1265 #endif // SPARKS
1266 }
1267 #endif // PARALLEL_HASKELL
1268
1269 /* ----------------------------------------------------------------------------
1270 * Get work from a remote node (PARALLEL_HASKELL only)
1271 * ------------------------------------------------------------------------- */
1272
1273 #if defined(PARALLEL_HASKELL)
1274 static rtsBool
1275 scheduleGetRemoteWork(rtsBool *receivedFinish)
1276 {
1277 ASSERT(emptyRunQueue());
1278
1279 if (RtsFlags.ParFlags.BufferTime) {
1280 IF_PAR_DEBUG(verbose,
1281 debugBelch("...send all pending data,"));
1282 {
1283 nat i;
1284 for (i=1; i<=nPEs; i++)
1285 sendImmediately(i); // send all messages away immediately
1286 }
1287 }
1288 # ifndef SPARKS
1289 //++EDEN++ idle() , i.e. send all buffers, wait for work
1290 // suppress fishing in EDEN... just look for incoming messages
1291 // (blocking receive)
1292 IF_PAR_DEBUG(verbose,
1293 debugBelch("...wait for incoming messages...\n"));
1294 *receivedFinish = processMessages(); // blocking receive...
1295
1296 // and reenter scheduling loop after having received something
1297 // (return rtsFalse below)
1298
1299 # else /* activate SPARKS machinery */
1300 /* We get here, if we have no work, tried to activate a local spark, but still
1301 have no work. We try to get a remote spark, by sending a FISH message.
1302 Thread migration should be added here, and triggered when a sequence of
1303 fishes returns without work. */
1304 delay = (RtsFlags.ParFlags.fishDelay!=0ll ? RtsFlags.ParFlags.fishDelay : 0ll);
1305
1306 /* =8-[ no local sparks => look for work on other PEs */
1307 /*
1308 * We really have absolutely no work. Send out a fish
1309 * (there may be some out there already), and wait for
1310 * something to arrive. We clearly can't run any threads
1311 * until a SCHEDULE or RESUME arrives, and so that's what
1312 * we're hoping to see. (Of course, we still have to
1313 * respond to other types of messages.)
1314 */
1315 rtsTime now = msTime() /*CURRENT_TIME*/;
1316 IF_PAR_DEBUG(verbose,
1317 debugBelch("-- now=%ld\n", now));
1318 IF_PAR_DEBUG(fish, // verbose,
1319 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
1320 (last_fish_arrived_at!=0 &&
1321 last_fish_arrived_at+delay > now)) {
1322 debugBelch("--$$ <%llu> delaying FISH until %llu (last fish %llu, delay %llu)\n",
1323 now, last_fish_arrived_at+delay,
1324 last_fish_arrived_at,
1325 delay);
1326 });
1327
1328 if (outstandingFishes < RtsFlags.ParFlags.maxFishes &&
1329 advisory_thread_count < RtsFlags.ParFlags.maxThreads) { // send a FISH, but when?
1330 if (last_fish_arrived_at==0 ||
1331 (last_fish_arrived_at+delay <= now)) { // send FISH now!
1332 /* outstandingFishes is set in sendFish, processFish;
1333 avoid flooding system with fishes via delay */
1334 next_fish_to_send_at = 0;
1335 } else {
1336 /* ToDo: this should be done in the main scheduling loop to avoid the
1337 busy wait here; not so bad if fish delay is very small */
1338 int iq = 0; // DEBUGGING -- HWL
1339 next_fish_to_send_at = last_fish_arrived_at+delay; // remember when to send
1340 /* send a fish when ready, but process messages that arrive in the meantime */
1341 do {
1342 if (PacketsWaiting()) {
1343 iq++; // DEBUGGING
1344 *receivedFinish = processMessages();
1345 }
1346 now = msTime();
1347 } while (!*receivedFinish || now<next_fish_to_send_at);
1348 // JB: This means the fish could become obsolete, if we receive
1349 // work. Better check for work again?
1350 // last line: while (!receivedFinish || !haveWork || now<...)
1351 // next line: if (receivedFinish || haveWork )
1352
1353 if (*receivedFinish) // no need to send a FISH if we are finishing anyway
1354 return rtsFalse; // NB: this will leave scheduler loop
1355 // immediately after return!
1356
1357 IF_PAR_DEBUG(fish, // verbose,
1358 debugBelch("--$$ <%llu> sent delayed fish (%d processMessages); active/total threads=%d/%d\n",now,iq,run_queue_len(),advisory_thread_count));
1359
1360 }
1361
1362 // JB: IMHO, this should all be hidden inside sendFish(...)
1363 /* pe = choosePE();
1364 sendFish(pe, thisPE, NEW_FISH_AGE, NEW_FISH_HISTORY,
1365 NEW_FISH_HUNGER);
1366
1367 // Global statistics: count no. of fishes
1368 if (RtsFlags.ParFlags.ParStats.Global &&
1369 RtsFlags.GcFlags.giveStats > NO_GC_STATS) {
1370 globalParStats.tot_fish_mess++;
1371 }
1372 */
1373
1374 /* delayed fishes must have been sent by now! */
1375 next_fish_to_send_at = 0;
1376 }
1377
1378 *receivedFinish = processMessages();
1379 # endif /* SPARKS */
1380
1381 return rtsFalse;
1382 /* NB: this function always returns rtsFalse, meaning the scheduler
1383 loop continues with the next iteration;
1384 rationale:
1385 return code means success in finding work; we enter this function
1386 if there is no local work, thus have to send a fish which takes
1387 time until it arrives with work; in the meantime we should process
1388 messages in the main loop;
1389 */
1390 }
1391 #endif // PARALLEL_HASKELL
1392
1393 /* ----------------------------------------------------------------------------
1394 * PAR/GRAN: Report stats & debugging info(?)
1395 * ------------------------------------------------------------------------- */
1396
1397 #if defined(PAR) || defined(GRAN)
1398 static void
1399 scheduleGranParReport(void)
1400 {
1401 ASSERT(run_queue_hd != END_TSO_QUEUE);
1402
1403 /* Take a thread from the run queue, if we have work */
1404 POP_RUN_QUEUE(t); // take_off_run_queue(END_TSO_QUEUE);
1405
1406 /* If this TSO has got its outport closed in the meantime,
1407 * it mustn't be run. Instead, we have to clean it up as if it was finished.
1408 * It has to be marked as TH_DEAD for this purpose.
1409 * If it is TH_TERM instead, it is supposed to have finished in the normal way.
1410
1411 JB: TODO: investigate wether state change field could be nuked
1412 entirely and replaced by the normal tso state (whatnext
1413 field). All we want to do is to kill tsos from outside.
1414 */
1415
1416 /* ToDo: write something to the log-file
1417 if (RTSflags.ParFlags.granSimStats && !sameThread)
1418 DumpGranEvent(GR_SCHEDULE, RunnableThreadsHd);
1419
1420 CurrentTSO = t;
1421 */
1422 /* the spark pool for the current PE */
1423 pool = &(cap.r.rSparks); // cap = (old) MainCap
1424
1425 IF_DEBUG(scheduler,
1426 debugBelch("--=^ %d threads, %d sparks on [%#x]\n",
1427 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
1428
1429 IF_PAR_DEBUG(fish,
1430 debugBelch("--=^ %d threads, %d sparks on [%#x]\n",
1431 run_queue_len(), spark_queue_len(pool), CURRENT_PROC));
1432
1433 if (RtsFlags.ParFlags.ParStats.Full &&
1434 (t->par.sparkname != (StgInt)0) && // only log spark generated threads
1435 (emitSchedule || // forced emit
1436 (t && LastTSO && t->id != LastTSO->id))) {
1437 /*
1438 we are running a different TSO, so write a schedule event to log file
1439 NB: If we use fair scheduling we also have to write a deschedule
1440 event for LastTSO; with unfair scheduling we know that the
1441 previous tso has blocked whenever we switch to another tso, so
1442 we don't need it in GUM for now
1443 */
1444 IF_PAR_DEBUG(fish, // schedule,
1445 debugBelch("____ scheduling spark generated thread %d (%lx) (%lx) via a forced emit\n",t->id,t,t->par.sparkname));
1446
1447 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1448 GR_SCHEDULE, t, (StgClosure *)NULL, 0, 0);
1449 emitSchedule = rtsFalse;
1450 }
1451 }
1452 #endif
1453
1454 /* ----------------------------------------------------------------------------
1455 * After running a thread...
1456 * ------------------------------------------------------------------------- */
1457
1458 static void
1459 schedulePostRunThread(void)
1460 {
1461 #if defined(PAR)
1462 /* HACK 675: if the last thread didn't yield, make sure to print a
1463 SCHEDULE event to the log file when StgRunning the next thread, even
1464 if it is the same one as before */
1465 LastTSO = t;
1466 TimeOfLastYield = CURRENT_TIME;
1467 #endif
1468
1469 /* some statistics gathering in the parallel case */
1470
1471 #if defined(GRAN) || defined(PAR) || defined(EDEN)
1472 switch (ret) {
1473 case HeapOverflow:
1474 # if defined(GRAN)
1475 IF_DEBUG(gran, DumpGranEvent(GR_DESCHEDULE, t));
1476 globalGranStats.tot_heapover++;
1477 # elif defined(PAR)
1478 globalParStats.tot_heapover++;
1479 # endif
1480 break;
1481
1482 case StackOverflow:
1483 # if defined(GRAN)
1484 IF_DEBUG(gran,
1485 DumpGranEvent(GR_DESCHEDULE, t));
1486 globalGranStats.tot_stackover++;
1487 # elif defined(PAR)
1488 // IF_DEBUG(par,
1489 // DumpGranEvent(GR_DESCHEDULE, t);
1490 globalParStats.tot_stackover++;
1491 # endif
1492 break;
1493
1494 case ThreadYielding:
1495 # if defined(GRAN)
1496 IF_DEBUG(gran,
1497 DumpGranEvent(GR_DESCHEDULE, t));
1498 globalGranStats.tot_yields++;
1499 # elif defined(PAR)
1500 // IF_DEBUG(par,
1501 // DumpGranEvent(GR_DESCHEDULE, t);
1502 globalParStats.tot_yields++;
1503 # endif
1504 break;
1505
1506 case ThreadBlocked:
1507 # if defined(GRAN)
1508 debugTrace(DEBUG_sched,
1509 "--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: ",
1510 t->id, t, whatNext_strs[t->what_next], t->block_info.closure,
1511 (t->block_info.closure==(StgClosure*)NULL ? 99 : where_is(t->block_info.closure)));
1512 if (t->block_info.closure!=(StgClosure*)NULL)
1513 print_bq(t->block_info.closure);
1514 debugBelch("\n"));
1515
1516 // ??? needed; should emit block before
1517 IF_DEBUG(gran,
1518 DumpGranEvent(GR_DESCHEDULE, t));
1519 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1520 /*
1521 ngoq Dogh!
1522 ASSERT(procStatus[CurrentProc]==Busy ||
1523 ((procStatus[CurrentProc]==Fetching) &&
1524 (t->block_info.closure!=(StgClosure*)NULL)));
1525 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1526 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1527 procStatus[CurrentProc]==Fetching))
1528 procStatus[CurrentProc] = Idle;
1529 */
1530 # elif defined(PAR)
1531 //++PAR++ blockThread() writes the event (change?)
1532 # endif
1533 break;
1534
1535 case ThreadFinished:
1536 break;
1537
1538 default:
1539 barf("parGlobalStats: unknown return code");
1540 break;
1541 }
1542 #endif
1543 }
1544
1545 /* -----------------------------------------------------------------------------
1546 * Handle a thread that returned to the scheduler with ThreadHeepOverflow
1547 * -------------------------------------------------------------------------- */
1548
1549 static rtsBool
1550 scheduleHandleHeapOverflow( Capability *cap, StgTSO *t )
1551 {
1552 // did the task ask for a large block?
1553 if (cap->r.rHpAlloc > BLOCK_SIZE) {
1554 // if so, get one and push it on the front of the nursery.
1555 bdescr *bd;
1556 lnat blocks;
1557
1558 blocks = (lnat)BLOCK_ROUND_UP(cap->r.rHpAlloc) / BLOCK_SIZE;
1559
1560 debugTrace(DEBUG_sched,
1561 "--<< thread %ld (%s) stopped: requesting a large block (size %ld)\n",
1562 (long)t->id, whatNext_strs[t->what_next], blocks);
1563
1564 // don't do this if the nursery is (nearly) full, we'll GC first.
1565 if (cap->r.rCurrentNursery->link != NULL ||
1566 cap->r.rNursery->n_blocks == 1) { // paranoia to prevent infinite loop
1567 // if the nursery has only one block.
1568
1569 ACQUIRE_SM_LOCK
1570 bd = allocGroup( blocks );
1571 RELEASE_SM_LOCK
1572 cap->r.rNursery->n_blocks += blocks;
1573
1574 // link the new group into the list
1575 bd->link = cap->r.rCurrentNursery;
1576 bd->u.back = cap->r.rCurrentNursery->u.back;
1577 if (cap->r.rCurrentNursery->u.back != NULL) {
1578 cap->r.rCurrentNursery->u.back->link = bd;
1579 } else {
1580 #if !defined(THREADED_RTS)
1581 ASSERT(g0s0->blocks == cap->r.rCurrentNursery &&
1582 g0s0 == cap->r.rNursery);
1583 #endif
1584 cap->r.rNursery->blocks = bd;
1585 }
1586 cap->r.rCurrentNursery->u.back = bd;
1587
1588 // initialise it as a nursery block. We initialise the
1589 // step, gen_no, and flags field of *every* sub-block in
1590 // this large block, because this is easier than making
1591 // sure that we always find the block head of a large
1592 // block whenever we call Bdescr() (eg. evacuate() and
1593 // isAlive() in the GC would both have to do this, at
1594 // least).
1595 {
1596 bdescr *x;
1597 for (x = bd; x < bd + blocks; x++) {
1598 x->step = cap->r.rNursery;
1599 x->gen_no = 0;
1600 x->flags = 0;
1601 }
1602 }
1603
1604 // This assert can be a killer if the app is doing lots
1605 // of large block allocations.
1606 IF_DEBUG(sanity, checkNurserySanity(cap->r.rNursery));
1607
1608 // now update the nursery to point to the new block
1609 cap->r.rCurrentNursery = bd;
1610
1611 // we might be unlucky and have another thread get on the
1612 // run queue before us and steal the large block, but in that
1613 // case the thread will just end up requesting another large
1614 // block.
1615 pushOnRunQueue(cap,t);
1616 return rtsFalse; /* not actually GC'ing */
1617 }
1618 }
1619
1620 debugTrace(DEBUG_sched,
1621 "--<< thread %ld (%s) stopped: HeapOverflow\n",
1622 (long)t->id, whatNext_strs[t->what_next]);
1623
1624 #if defined(GRAN)
1625 ASSERT(!is_on_queue(t,CurrentProc));
1626 #elif defined(PARALLEL_HASKELL)
1627 /* Currently we emit a DESCHEDULE event before GC in GUM.
1628 ToDo: either add separate event to distinguish SYSTEM time from rest
1629 or just nuke this DESCHEDULE (and the following SCHEDULE) */
1630 if (0 && RtsFlags.ParFlags.ParStats.Full) {
1631 DumpRawGranEvent(CURRENT_PROC, CURRENT_PROC,
1632 GR_DESCHEDULE, t, (StgClosure *)NULL, 0, 0);
1633 emitSchedule = rtsTrue;
1634 }
1635 #endif
1636
1637 pushOnRunQueue(cap,t);
1638 return rtsTrue;
1639 /* actual GC is done at the end of the while loop in schedule() */
1640 }
1641
1642 /* -----------------------------------------------------------------------------
1643 * Handle a thread that returned to the scheduler with ThreadStackOverflow
1644 * -------------------------------------------------------------------------- */
1645
1646 static void
1647 scheduleHandleStackOverflow (Capability *cap, Task *task, StgTSO *t)
1648 {
1649 debugTrace (DEBUG_sched,
1650 "--<< thread %ld (%s) stopped, StackOverflow",
1651 (long)t->id, whatNext_strs[t->what_next]);
1652
1653 /* just adjust the stack for this thread, then pop it back
1654 * on the run queue.
1655 */
1656 {
1657 /* enlarge the stack */
1658 StgTSO *new_t = threadStackOverflow(cap, t);
1659
1660 /* The TSO attached to this Task may have moved, so update the
1661 * pointer to it.
1662 */
1663 if (task->tso == t) {
1664 task->tso = new_t;
1665 }
1666 pushOnRunQueue(cap,new_t);
1667 }
1668 }
1669
1670 /* -----------------------------------------------------------------------------
1671 * Handle a thread that returned to the scheduler with ThreadYielding
1672 * -------------------------------------------------------------------------- */
1673
1674 static rtsBool
1675 scheduleHandleYield( Capability *cap, StgTSO *t, nat prev_what_next )
1676 {
1677 // Reset the context switch flag. We don't do this just before
1678 // running the thread, because that would mean we would lose ticks
1679 // during GC, which can lead to unfair scheduling (a thread hogs
1680 // the CPU because the tick always arrives during GC). This way
1681 // penalises threads that do a lot of allocation, but that seems
1682 // better than the alternative.
1683 context_switch = 0;
1684
1685 /* put the thread back on the run queue. Then, if we're ready to
1686 * GC, check whether this is the last task to stop. If so, wake
1687 * up the GC thread. getThread will block during a GC until the
1688 * GC is finished.
1689 */
1690 #ifdef DEBUG
1691 if (t->what_next != prev_what_next) {
1692 debugTrace(DEBUG_sched,
1693 "--<< thread %ld (%s) stopped to switch evaluators",
1694 (long)t->id, whatNext_strs[t->what_next]);
1695 } else {
1696 debugTrace(DEBUG_sched,
1697 "--<< thread %ld (%s) stopped, yielding",
1698 (long)t->id, whatNext_strs[t->what_next]);
1699 }
1700 #endif
1701
1702 IF_DEBUG(sanity,
1703 //debugBelch("&& Doing sanity check on yielding TSO %ld.", t->id);
1704 checkTSO(t));
1705 ASSERT(t->link == END_TSO_QUEUE);
1706
1707 // Shortcut if we're just switching evaluators: don't bother
1708 // doing stack squeezing (which can be expensive), just run the
1709 // thread.
1710 if (t->what_next != prev_what_next) {
1711 return rtsTrue;
1712 }
1713
1714 #if defined(GRAN)
1715 ASSERT(!is_on_queue(t,CurrentProc));
1716
1717 IF_DEBUG(sanity,
1718 //debugBelch("&& Doing sanity check on all ThreadQueues (and their TSOs).");
1719 checkThreadQsSanity(rtsTrue));
1720
1721 #endif
1722
1723 addToRunQueue(cap,t);
1724
1725 #if defined(GRAN)
1726 /* add a ContinueThread event to actually process the thread */
1727 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
1728 ContinueThread,
1729 t, (StgClosure*)NULL, (rtsSpark*)NULL);
1730 IF_GRAN_DEBUG(bq,
1731 debugBelch("GRAN: eventq and runnableq after adding yielded thread to queue again:\n");
1732 G_EVENTQ(0);
1733 G_CURR_THREADQ(0));
1734 #endif
1735 return rtsFalse;
1736 }
1737
1738 /* -----------------------------------------------------------------------------
1739 * Handle a thread that returned to the scheduler with ThreadBlocked
1740 * -------------------------------------------------------------------------- */
1741
1742 static void
1743 scheduleHandleThreadBlocked( StgTSO *t
1744 #if !defined(GRAN) && !defined(DEBUG)
1745 STG_UNUSED
1746 #endif
1747 )
1748 {
1749 #if defined(GRAN)
1750 IF_DEBUG(scheduler,
1751 debugBelch("--<< thread %ld (%p; %s) stopped, blocking on node %p [PE %d] with BQ: \n",
1752 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)));
1753 if (t->block_info.closure!=(StgClosure*)NULL) print_bq(t->block_info.closure));
1754
1755 // ??? needed; should emit block before
1756 IF_DEBUG(gran,
1757 DumpGranEvent(GR_DESCHEDULE, t));
1758 prune_eventq(t, (StgClosure *)NULL); // prune ContinueThreads for t
1759 /*
1760 ngoq Dogh!
1761 ASSERT(procStatus[CurrentProc]==Busy ||
1762 ((procStatus[CurrentProc]==Fetching) &&
1763 (t->block_info.closure!=(StgClosure*)NULL)));
1764 if (run_queue_hds[CurrentProc] == END_TSO_QUEUE &&
1765 !(!RtsFlags.GranFlags.DoAsyncFetch &&
1766 procStatus[CurrentProc]==Fetching))
1767 procStatus[CurrentProc] = Idle;
1768 */
1769 #elif defined(PAR)
1770 IF_DEBUG(scheduler,
1771 debugBelch("--<< thread %ld (%p; %s) stopped, blocking on node %p with BQ: \n",
1772 t->id, t, whatNext_strs[t->what_next], t->block_info.closure));
1773 IF_PAR_DEBUG(bq,
1774
1775 if (t->block_info.closure!=(StgClosure*)NULL)
1776 print_bq(t->block_info.closure));
1777
1778 /* Send a fetch (if BlockedOnGA) and dump event to log file */
1779 blockThread(t);
1780
1781 /* whatever we schedule next, we must log that schedule */
1782 emitSchedule = rtsTrue;
1783
1784 #else /* !GRAN */
1785
1786 // We don't need to do anything. The thread is blocked, and it
1787 // has tidied up its stack and placed itself on whatever queue
1788 // it needs to be on.
1789
1790 // ASSERT(t->why_blocked != NotBlocked);
1791 // Not true: for example,
1792 // - in THREADED_RTS, the thread may already have been woken
1793 // up by another Capability. This actually happens: try
1794 // conc023 +RTS -N2.
1795 // - the thread may have woken itself up already, because
1796 // threadPaused() might have raised a blocked throwTo
1797 // exception, see maybePerformBlockedException().
1798
1799 #ifdef DEBUG
1800 if (traceClass(DEBUG_sched)) {
1801 debugTraceBegin("--<< thread %lu (%s) stopped: ",
1802 (unsigned long)t->id, whatNext_strs[t->what_next]);
1803 printThreadBlockage(t);
1804 debugTraceEnd();
1805 }
1806 #endif
1807
1808 /* Only for dumping event to log file
1809 ToDo: do I need this in GranSim, too?
1810 blockThread(t);
1811 */
1812 #endif
1813 }
1814
1815 /* -----------------------------------------------------------------------------
1816 * Handle a thread that returned to the scheduler with ThreadFinished
1817 * -------------------------------------------------------------------------- */
1818
1819 static rtsBool
1820 scheduleHandleThreadFinished (Capability *cap STG_UNUSED, Task *task, StgTSO *t)
1821 {
1822 /* Need to check whether this was a main thread, and if so,
1823 * return with the return value.
1824 *
1825 * We also end up here if the thread kills itself with an
1826 * uncaught exception, see Exception.cmm.
1827 */
1828 debugTrace(DEBUG_sched, "--++ thread %lu (%s) finished",
1829 (unsigned long)t->id, whatNext_strs[t->what_next]);
1830
1831 #if defined(GRAN)
1832 endThread(t, CurrentProc); // clean-up the thread
1833 #elif defined(PARALLEL_HASKELL)
1834 /* For now all are advisory -- HWL */
1835 //if(t->priority==AdvisoryPriority) ??
1836 advisory_thread_count--; // JB: Caution with this counter, buggy!
1837
1838 # if defined(DIST)
1839 if(t->dist.priority==RevalPriority)
1840 FinishReval(t);
1841 # endif
1842
1843 # if defined(EDENOLD)
1844 // the thread could still have an outport... (BUG)
1845 if (t->eden.outport != -1) {
1846 // delete the outport for the tso which has finished...
1847 IF_PAR_DEBUG(eden_ports,
1848 debugBelch("WARNING: Scheduler removes outport %d for TSO %d.\n",
1849 t->eden.outport, t->id));
1850 deleteOPT(t);
1851 }
1852 // thread still in the process (HEAVY BUG! since outport has just been closed...)
1853 if (t->eden.epid != -1) {
1854 IF_PAR_DEBUG(eden_ports,
1855 debugBelch("WARNING: Scheduler removes TSO %d from process %d .\n",
1856 t->id, t->eden.epid));
1857 removeTSOfromProcess(t);
1858 }
1859 # endif
1860
1861 # if defined(PAR)
1862 if (RtsFlags.ParFlags.ParStats.Full &&
1863 !RtsFlags.ParFlags.ParStats.Suppressed)
1864 DumpEndEvent(CURRENT_PROC, t, rtsFalse /* not mandatory */);
1865
1866 // t->par only contains statistics: left out for now...
1867 IF_PAR_DEBUG(fish,
1868 debugBelch("**** end thread: ended sparked thread %d (%lx); sparkname: %lx\n",
1869 t->id,t,t->par.sparkname));
1870 # endif
1871 #endif // PARALLEL_HASKELL
1872
1873 //
1874 // Check whether the thread that just completed was a bound
1875 // thread, and if so return with the result.
1876 //
1877 // There is an assumption here that all thread completion goes
1878 // through this point; we need to make sure that if a thread
1879 // ends up in the ThreadKilled state, that it stays on the run
1880 // queue so it can be dealt with here.
1881 //
1882
1883 if (t->bound) {
1884
1885 if (t->bound != task) {
1886 #if !defined(THREADED_RTS)
1887 // Must be a bound thread that is not the topmost one. Leave
1888 // it on the run queue until the stack has unwound to the
1889 // point where we can deal with this. Leaving it on the run
1890 // queue also ensures that the garbage collector knows about
1891 // this thread and its return value (it gets dropped from the
1892 // all_threads list so there's no other way to find it).
1893 appendToRunQueue(cap,t);
1894 return rtsFalse;
1895 #else
1896 // this cannot happen in the threaded RTS, because a
1897 // bound thread can only be run by the appropriate Task.
1898 barf("finished bound thread that isn't mine");
1899 #endif
1900 }
1901
1902 ASSERT(task->tso == t);
1903
1904 if (t->what_next == ThreadComplete) {
1905 if (task->ret) {
1906 // NOTE: return val is tso->sp[1] (see StgStartup.hc)
1907 *(task->ret) = (StgClosure *)task->tso->sp[1];
1908 }
1909 task->stat = Success;
1910 } else {
1911 if (task->ret) {
1912 *(task->ret) = NULL;
1913 }
1914 if (sched_state >= SCHED_INTERRUPTING) {
1915 task->stat = Interrupted;
1916 } else {
1917 task->stat = Killed;
1918 }
1919 }
1920 #ifdef DEBUG
1921 removeThreadLabel((StgWord)task->tso->id);
1922 #endif
1923 return rtsTrue; // tells schedule() to return
1924 }
1925
1926 return rtsFalse;
1927 }
1928
1929 /* -----------------------------------------------------------------------------
1930 * Perform a heap census
1931 * -------------------------------------------------------------------------- */
1932
1933 static rtsBool
1934 scheduleNeedHeapProfile( rtsBool ready_to_gc STG_UNUSED )
1935 {
1936 // When we have +RTS -i0 and we're heap profiling, do a census at
1937 // every GC. This lets us get repeatable runs for debugging.
1938 if (performHeapProfile ||
1939 (RtsFlags.ProfFlags.profileInterval==0 &&
1940 RtsFlags.ProfFlags.doHeapProfile && ready_to_gc)) {
1941 return rtsTrue;
1942 } else {
1943 return rtsFalse;
1944 }
1945 }
1946
1947 /* -----------------------------------------------------------------------------
1948 * Perform a garbage collection if necessary
1949 * -------------------------------------------------------------------------- */
1950
1951 static Capability *
1952 scheduleDoGC (Capability *cap, Task *task USED_IF_THREADS, rtsBool force_major)
1953 {
1954 StgTSO *t;
1955 rtsBool heap_census;
1956 #ifdef THREADED_RTS
1957 static volatile StgWord waiting_for_gc;
1958 rtsBool was_waiting;
1959 nat i;
1960 #endif
1961
1962 #ifdef THREADED_RTS
1963 // In order to GC, there must be no threads running Haskell code.
1964 // Therefore, the GC thread needs to hold *all* the capabilities,
1965 // and release them after the GC has completed.
1966 //
1967 // This seems to be the simplest way: previous attempts involved
1968 // making all the threads with capabilities give up their
1969 // capabilities and sleep except for the *last* one, which
1970 // actually did the GC. But it's quite hard to arrange for all
1971 // the other tasks to sleep and stay asleep.
1972 //
1973
1974 was_waiting = cas(&waiting_for_gc, 0, 1);
1975 if (was_waiting) {
1976 do {
1977 debugTrace(DEBUG_sched, "someone else is trying to GC...");
1978 if (cap) yieldCapability(&cap,task);
1979 } while (waiting_for_gc);
1980 return cap; // NOTE: task->cap might have changed here
1981 }
1982
1983 for (i=0; i < n_capabilities; i++) {
1984 debugTrace(DEBUG_sched, "ready_to_gc, grabbing all the capabilies (%d/%d)", i, n_capabilities);
1985 if (cap != &capabilities[i]) {
1986 Capability *pcap = &capabilities[i];
1987 // we better hope this task doesn't get migrated to
1988 // another Capability while we're waiting for this one.
1989 // It won't, because load balancing happens while we have
1990 // all the Capabilities, but even so it's a slightly
1991 // unsavoury invariant.
1992 task->cap = pcap;
1993 context_switch = 1;
1994 waitForReturnCapability(&pcap, task);
1995 if (pcap != &capabilities[i]) {
1996 barf("scheduleDoGC: got the wrong capability");
1997 }
1998 }
1999 }
2000
2001 waiting_for_gc = rtsFalse;
2002 #endif
2003
2004 /* Kick any transactions which are invalid back to their
2005 * atomically frames. When next scheduled they will try to
2006 * commit, this commit will fail and they will retry.
2007 */
2008 {
2009 StgTSO *next;
2010
2011 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
2012 if (t->what_next == ThreadRelocated) {
2013 next = t->link;
2014 } else {
2015 next = t->global_link;
2016
2017 // This is a good place to check for blocked
2018 // exceptions. It might be the case that a thread is
2019 // blocked on delivering an exception to a thread that
2020 // is also blocked - we try to ensure that this
2021 // doesn't happen in throwTo(), but it's too hard (or
2022 // impossible) to close all the race holes, so we
2023 // accept that some might get through and deal with
2024 // them here. A GC will always happen at some point,
2025 // even if the system is otherwise deadlocked.
2026 maybePerformBlockedException (&capabilities[0], t);
2027
2028 if (t -> trec != NO_TREC && t -> why_blocked == NotBlocked) {
2029 if (!stmValidateNestOfTransactions (t -> trec)) {
2030 debugTrace(DEBUG_sched | DEBUG_stm,
2031 "trec %p found wasting its time", t);
2032
2033 // strip the stack back to the
2034 // ATOMICALLY_FRAME, aborting the (nested)
2035 // transaction, and saving the stack of any
2036 // partially-evaluated thunks on the heap.
2037 throwToSingleThreaded_(&capabilities[0], t,
2038 NULL, rtsTrue, NULL);
2039
2040 #ifdef REG_R1
2041 ASSERT(get_itbl((StgClosure *)t->sp)->type == ATOMICALLY_FRAME);
2042 #endif
2043 }
2044 }
2045 }
2046 }
2047 }
2048
2049 // so this happens periodically:
2050 if (cap) scheduleCheckBlackHoles(cap);
2051
2052 IF_DEBUG(scheduler, printAllThreads());
2053
2054 /*
2055 * We now have all the capabilities; if we're in an interrupting
2056 * state, then we should take the opportunity to delete all the
2057 * threads in the system.
2058 */
2059 if (sched_state >= SCHED_INTERRUPTING) {
2060 deleteAllThreads(&capabilities[0]);
2061 sched_state = SCHED_SHUTTING_DOWN;
2062 }
2063
2064 heap_census = scheduleNeedHeapProfile(rtsTrue);
2065
2066 /* everybody back, start the GC.
2067 * Could do it in this thread, or signal a condition var
2068 * to do it in another thread. Either way, we need to
2069 * broadcast on gc_pending_cond afterward.
2070 */
2071 #if defined(THREADED_RTS)
2072 debugTrace(DEBUG_sched, "doing GC");
2073 #endif
2074 GarbageCollect(force_major || heap_census);
2075
2076 if (heap_census) {
2077 debugTrace(DEBUG_sched, "performing heap census");
2078 heapCensus();
2079 performHeapProfile = rtsFalse;
2080 }
2081
2082 #if defined(THREADED_RTS)
2083 // release our stash of capabilities.
2084 for (i = 0; i < n_capabilities; i++) {
2085 if (cap != &capabilities[i]) {
2086 task->cap = &capabilities[i];
2087 releaseCapability(&capabilities[i]);
2088 }
2089 }
2090 if (cap) {
2091 task->cap = cap;
2092 } else {
2093 task->cap = NULL;
2094 }
2095 #endif
2096
2097 #if defined(GRAN)
2098 /* add a ContinueThread event to continue execution of current thread */
2099 new_event(CurrentProc, CurrentProc, CurrentTime[CurrentProc],
2100 ContinueThread,
2101 t, (StgClosure*)NULL, (rtsSpark*)NULL);
2102 IF_GRAN_DEBUG(bq,
2103 debugBelch("GRAN: eventq and runnableq after Garbage collection:\n\n");
2104 G_EVENTQ(0);
2105 G_CURR_THREADQ(0));
2106 #endif /* GRAN */
2107
2108 return cap;
2109 }
2110
2111 /* ---------------------------------------------------------------------------
2112 * Singleton fork(). Do not copy any running threads.
2113 * ------------------------------------------------------------------------- */
2114
2115 pid_t
2116 forkProcess(HsStablePtr *entry
2117 #ifndef FORKPROCESS_PRIMOP_SUPPORTED
2118 STG_UNUSED
2119 #endif
2120 )
2121 {
2122 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
2123 Task *task;
2124 pid_t pid;
2125 StgTSO* t,*next;
2126 Capability *cap;
2127
2128 #if defined(THREADED_RTS)
2129 if (RtsFlags.ParFlags.nNodes > 1) {
2130 errorBelch("forking not supported with +RTS -N<n> greater than 1");
2131 stg_exit(EXIT_FAILURE);
2132 }
2133 #endif
2134
2135 debugTrace(DEBUG_sched, "forking!");
2136
2137 // ToDo: for SMP, we should probably acquire *all* the capabilities
2138 cap = rts_lock();
2139
2140 // no funny business: hold locks while we fork, otherwise if some
2141 // other thread is holding a lock when the fork happens, the data
2142 // structure protected by the lock will forever be in an
2143 // inconsistent state in the child. See also #1391.
2144 ACQUIRE_LOCK(&sched_mutex);
2145 ACQUIRE_LOCK(&cap->lock);
2146 ACQUIRE_LOCK(&cap->running_task->lock);
2147
2148 pid = fork();
2149
2150 if (pid) { // parent
2151
2152 RELEASE_LOCK(&sched_mutex);
2153 RELEASE_LOCK(&cap->lock);
2154 RELEASE_LOCK(&cap->running_task->lock);
2155
2156 // just return the pid
2157 rts_unlock(cap);
2158 return pid;
2159
2160 } else { // child
2161
2162 #if defined(THREADED_RTS)
2163 initMutex(&sched_mutex);
2164 initMutex(&cap->lock);
2165 initMutex(&cap->running_task->lock);
2166 #endif
2167
2168 // Now, all OS threads except the thread that forked are
2169 // stopped. We need to stop all Haskell threads, including
2170 // those involved in foreign calls. Also we need to delete
2171 // all Tasks, because they correspond to OS threads that are
2172 // now gone.
2173
2174 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
2175 if (t->what_next == ThreadRelocated) {
2176 next = t->link;
2177 } else {
2178 next = t->global_link;
2179 // don't allow threads to catch the ThreadKilled
2180 // exception, but we do want to raiseAsync() because these
2181 // threads may be evaluating thunks that we need later.
2182 deleteThread_(cap,t);
2183 }
2184 }
2185
2186 // Empty the run queue. It seems tempting to let all the
2187 // killed threads stay on the run queue as zombies to be
2188 // cleaned up later, but some of them correspond to bound
2189 // threads for which the corresponding Task does not exist.
2190 cap->run_queue_hd = END_TSO_QUEUE;
2191 cap->run_queue_tl = END_TSO_QUEUE;
2192
2193 // Any suspended C-calling Tasks are no more, their OS threads
2194 // don't exist now:
2195 cap->suspended_ccalling_tasks = NULL;
2196
2197 // Empty the all_threads list. Otherwise, the garbage
2198 // collector may attempt to resurrect some of these threads.
2199 all_threads = END_TSO_QUEUE;
2200
2201 // Wipe the task list, except the current Task.
2202 ACQUIRE_LOCK(&sched_mutex);
2203 for (task = all_tasks; task != NULL; task=task->all_link) {
2204 if (task != cap->running_task) {
2205 #if defined(THREADED_RTS)
2206 initMutex(&task->lock); // see #1391
2207 #endif
2208 discardTask(task);
2209 }
2210 }
2211 RELEASE_LOCK(&sched_mutex);
2212
2213 #if defined(THREADED_RTS)
2214 // Wipe our spare workers list, they no longer exist. New
2215 // workers will be created if necessary.
2216 cap->spare_workers = NULL;
2217 cap->returning_tasks_hd = NULL;
2218 cap->returning_tasks_tl = NULL;
2219 #endif
2220
2221 // On Unix, all timers are reset in the child, so we need to start
2222 // the timer again.
2223 initTimer();
2224 startTimer();
2225
2226 cap = rts_evalStableIO(cap, entry, NULL); // run the action
2227 rts_checkSchedStatus("forkProcess",cap);
2228
2229 rts_unlock(cap);
2230 hs_exit(); // clean up and exit
2231 stg_exit(EXIT_SUCCESS);
2232 }
2233 #else /* !FORKPROCESS_PRIMOP_SUPPORTED */
2234 barf("forkProcess#: primop not supported on this platform, sorry!\n");
2235 return -1;
2236 #endif
2237 }
2238
2239 /* ---------------------------------------------------------------------------
2240 * Delete all the threads in the system
2241 * ------------------------------------------------------------------------- */
2242
2243 static void
2244 deleteAllThreads ( Capability *cap )
2245 {
2246 // NOTE: only safe to call if we own all capabilities.
2247
2248 StgTSO* t, *next;
2249 debugTrace(DEBUG_sched,"deleting all threads");
2250 for (t = all_threads; t != END_TSO_QUEUE; t = next) {
2251 if (t->what_next == ThreadRelocated) {
2252 next = t->link;
2253 } else {
2254 next = t->global_link;
2255 deleteThread(cap,t);
2256 }
2257 }
2258
2259 // The run queue now contains a bunch of ThreadKilled threads. We
2260 // must not throw these away: the main thread(s) will be in there
2261 // somewhere, and the main scheduler loop has to deal with it.
2262 // Also, the run queue is the only thing keeping these threads from
2263 // being GC'd, and we don't want the "main thread has been GC'd" panic.
2264
2265 #if !defined(THREADED_RTS)
2266 ASSERT(blocked_queue_hd == END_TSO_QUEUE);
2267 ASSERT(sleeping_queue == END_TSO_QUEUE);
2268 #endif
2269 }
2270
2271 /* -----------------------------------------------------------------------------
2272 Managing the suspended_ccalling_tasks list.
2273 Locks required: sched_mutex
2274 -------------------------------------------------------------------------- */
2275
2276 STATIC_INLINE void
2277 suspendTask (Capability *cap, Task *task)
2278 {
2279 ASSERT(task->next == NULL && task->prev == NULL);
2280 task->next = cap->suspended_ccalling_tasks;
2281 task->prev = NULL;
2282 if (cap->suspended_ccalling_tasks) {
2283 cap->suspended_ccalling_tasks->prev = task;
2284 }
2285 cap->suspended_ccalling_tasks = task;
2286 }
2287
2288 STATIC_INLINE void
2289 recoverSuspendedTask (Capability *cap, Task *task)
2290 {
2291 if (task->prev) {
2292 task->prev->next = task->next;
2293 } else {
2294 ASSERT(cap->suspended_ccalling_tasks == task);
2295 cap->suspended_ccalling_tasks = task->next;
2296 }
2297 if (task->next) {
2298 task->next->prev = task->prev;
2299 }
2300 task->next = task->prev = NULL;
2301 }
2302
2303 /* ---------------------------------------------------------------------------
2304 * Suspending & resuming Haskell threads.
2305 *
2306 * When making a "safe" call to C (aka _ccall_GC), the task gives back
2307 * its capability before calling the C function. This allows another
2308 * task to pick up the capability and carry on running Haskell
2309 * threads. It also means that if the C call blocks, it won't lock
2310 * the whole system.
2311 *
2312 * The Haskell thread making the C call is put to sleep for the
2313 * duration of the call, on the susepended_ccalling_threads queue. We
2314 * give out a token to the task, which it can use to resume the thread
2315 * on return from the C function.
2316 * ------------------------------------------------------------------------- */
2317
2318 void *
2319 suspendThread (StgRegTable *reg)
2320 {
2321 Capability *cap;
2322 int saved_errno;
2323 StgTSO *tso;
2324 Task *task;
2325 #if mingw32_HOST_OS
2326 StgWord32 saved_winerror;
2327 #endif
2328
2329 saved_errno = errno;
2330 #if mingw32_HOST_OS
2331 saved_winerror = GetLastError();
2332 #endif
2333
2334 /* assume that *reg is a pointer to the StgRegTable part of a Capability.
2335 */
2336 cap = regTableToCapability(reg);
2337
2338 task = cap->running_task;
2339 tso = cap->r.rCurrentTSO;
2340
2341 debugTrace(DEBUG_sched,
2342 "thread %lu did a safe foreign call",
2343 (unsigned long)cap->r.rCurrentTSO->id);
2344
2345 // XXX this might not be necessary --SDM
2346 tso->what_next = ThreadRunGHC;
2347
2348 threadPaused(cap,tso);
2349
2350 if ((tso->flags & TSO_BLOCKEX) == 0) {
2351 tso->why_blocked = BlockedOnCCall;
2352 tso->flags |= TSO_BLOCKEX;
2353 tso->flags &= ~TSO_INTERRUPTIBLE;
2354 } else {
2355 tso->why_blocked = BlockedOnCCall_NoUnblockExc;
2356 }
2357
2358 // Hand back capability
2359 task->suspended_tso = tso;
2360
2361 ACQUIRE_LOCK(&cap->lock);
2362
2363 suspendTask(cap,task);
2364 cap->in_haskell = rtsFalse;
2365 releaseCapability_(cap);
2366
2367 RELEASE_LOCK(&cap->lock);
2368
2369 #if defined(THREADED_RTS)
2370 /* Preparing to leave the RTS, so ensure there's a native thread/task
2371 waiting to take over.
2372 */
2373 debugTrace(DEBUG_sched, "thread %lu: leaving RTS", (unsigned long)tso->id);
2374 #endif
2375
2376 errno = saved_errno;
2377 #if mingw32_HOST_OS
2378 SetLastError(saved_winerror);
2379 #endif
2380 return task;
2381 }
2382
2383 StgRegTable *
2384 resumeThread (void *task_)
2385 {
2386 StgTSO *tso;
2387 Capability *cap;
2388 Task *task = task_;
2389 int saved_errno;
2390 #if mingw32_HOST_OS
2391 StgWord32 saved_winerror;
2392 #endif
2393
2394 saved_errno = errno;
2395 #if mingw32_HOST_OS
2396 saved_winerror = GetLastError();
2397 #endif
2398
2399 cap = task->cap;
2400 // Wait for permission to re-enter the RTS with the result.
2401 waitForReturnCapability(&cap,task);
2402 // we might be on a different capability now... but if so, our
2403 // entry on the suspended_ccalling_tasks list will also have been
2404 // migrated.
2405
2406 // Remove the thread from the suspended list
2407 recoverSuspendedTask(cap,task);
2408
2409 tso = task->suspended_tso;
2410 task->suspended_tso = NULL;
2411 tso->link = END_TSO_QUEUE;
2412 debugTrace(DEBUG_sched, "thread %lu: re-entering RTS", (unsigned long)tso->id);
2413
2414 if (tso->why_blocked == BlockedOnCCall) {
2415 awakenBlockedExceptionQueue(cap,tso);
2416 tso->flags &= ~(TSO_BLOCKEX | TSO_INTERRUPTIBLE);
2417 }
2418
2419 /* Reset blocking status */
2420 tso->why_blocked = NotBlocked;
2421
2422 cap->r.rCurrentTSO = tso;
2423 cap->in_haskell = rtsTrue;
2424 errno = saved_errno;
2425 #if mingw32_HOST_OS
2426 SetLastError(saved_winerror);
2427 #endif
2428
2429 /* We might have GC'd, mark the TSO dirty again */
2430 dirtyTSO(tso);
2431
2432 IF_DEBUG(sanity, checkTSO(tso));
2433
2434 return &cap->r;
2435 }
2436
2437 /* ---------------------------------------------------------------------------
2438 * scheduleThread()
2439 *
2440 * scheduleThread puts a thread on the end of the runnable queue.
2441 * This will usually be done immediately after a thread is created.
2442 * The caller of scheduleThread must create the thread using e.g.
2443 * createThread and push an appropriate closure
2444 * on this thread's stack before the scheduler is invoked.
2445 * ------------------------------------------------------------------------ */
2446
2447 void
2448 scheduleThread(Capability *cap, StgTSO *tso)
2449 {
2450 // The thread goes at the *end* of the run-queue, to avoid possible
2451 // starvation of any threads already on the queue.
2452 appendToRunQueue(cap,tso);
2453 }
2454
2455 void
2456 scheduleThreadOn(Capability *cap, StgWord cpu USED_IF_THREADS, StgTSO *tso)
2457 {
2458 #if defined(THREADED_RTS)
2459 tso->flags |= TSO_LOCKED; // we requested explicit affinity; don't
2460 // move this thread from now on.
2461 cpu %= RtsFlags.ParFlags.nNodes;
2462 if (cpu == cap->no) {
2463 appendToRunQueue(cap,tso);
2464 } else {
2465 migrateThreadToCapability_lock(&capabilities[cpu],tso);
2466 }
2467 #else
2468 appendToRunQueue(cap,tso);
2469 #endif
2470 }
2471
2472 Capability *
2473 scheduleWaitThread (StgTSO* tso, /*[out]*/HaskellObj* ret, Capability *cap)
2474 {
2475 Task *task;
2476
2477 // We already created/initialised the Task
2478 task = cap->running_task;
2479
2480 // This TSO is now a bound thread; make the Task and TSO
2481 // point to each other.
2482 tso->bound = task;
2483 tso->cap = cap;
2484
2485 task->tso = tso;
2486 task->ret = ret;
2487 task->stat = NoStatus;
2488
2489 appendToRunQueue(cap,tso);
2490
2491 debugTrace(DEBUG_sched, "new bound thread (%lu)", (unsigned long)tso->id);
2492
2493 #if defined(GRAN)
2494 /* GranSim specific init */
2495 CurrentTSO = m->tso; // the TSO to run
2496 procStatus[MainProc] = Busy; // status of main PE
2497 CurrentProc = MainProc; // PE to run it on
2498 #endif
2499
2500 cap = schedule(cap,task);
2501
2502 ASSERT(task->stat != NoStatus);
2503 ASSERT_FULL_CAPABILITY_INVARIANTS(cap,task);
2504
2505 debugTrace(DEBUG_sched, "bound thread (%lu) finished", (unsigned long)task->tso->id);
2506 return cap;
2507 }
2508
2509 /* ----------------------------------------------------------------------------
2510 * Starting Tasks
2511 * ------------------------------------------------------------------------- */
2512
2513 #if defined(THREADED_RTS)
2514 void
2515 workerStart(Task *task)
2516 {
2517 Capability *cap;
2518
2519 // See startWorkerTask().
2520 ACQUIRE_LOCK(&task->lock);
2521 cap = task->cap;
2522 RELEASE_LOCK(&task->lock);
2523
2524 // set the thread-local pointer to the Task:
2525 taskEnter(task);
2526
2527 // schedule() runs without a lock.
2528 cap = schedule(cap,task);
2529
2530 // On exit from schedule(), we have a Capability.
2531 releaseCapability(cap);
2532 workerTaskStop(task);
2533 }
2534 #endif
2535
2536 /* ---------------------------------------------------------------------------
2537 * initScheduler()
2538 *
2539 * Initialise the scheduler. This resets all the queues - if the
2540 * queues contained any threads, they'll be garbage collected at the
2541 * next pass.
2542 *
2543 * ------------------------------------------------------------------------ */
2544
2545 void
2546 initScheduler(void)
2547 {
2548 #if defined(GRAN)
2549 nat i;
2550 for (i=0; i<=MAX_PROC; i++) {
2551 run_queue_hds[i] = END_TSO_QUEUE;
2552 run_queue_tls[i] = END_TSO_QUEUE;
2553 blocked_queue_hds[i] = END_TSO_QUEUE;
2554 blocked_queue_tls[i] = END_TSO_QUEUE;
2555 ccalling_threadss[i] = END_TSO_QUEUE;
2556 blackhole_queue[i] = END_TSO_QUEUE;
2557 sleeping_queue = END_TSO_QUEUE;
2558 }
2559 #elif !defined(THREADED_RTS)
2560 blocked_queue_hd = END_TSO_QUEUE;
2561 blocked_queue_tl = END_TSO_QUEUE;
2562 sleeping_queue = END_TSO_QUEUE;
2563 #endif
2564
2565 blackhole_queue = END_TSO_QUEUE;
2566 all_threads = END_TSO_QUEUE;
2567
2568 context_switch = 0;
2569 sched_state = SCHED_RUNNING;
2570 recent_activity = ACTIVITY_YES;
2571
2572 #if defined(THREADED_RTS)
2573 /* Initialise the mutex and condition variables used by
2574 * the scheduler. */
2575 initMutex(&sched_mutex);
2576 #endif
2577
2578 ACQUIRE_LOCK(&sched_mutex);
2579
2580 /* A capability holds the state a native thread needs in
2581 * order to execute STG code. At least one capability is
2582 * floating around (only THREADED_RTS builds have more than one).
2583 */
2584 initCapabilities();
2585
2586 initTaskManager();
2587
2588 #if defined(THREADED_RTS) || defined(PARALLEL_HASKELL)
2589 initSparkPools();
2590 #endif
2591
2592 #if defined(THREADED_RTS)
2593 /*
2594 * Eagerly start one worker to run each Capability, except for
2595 * Capability 0. The idea is that we're probably going to start a
2596 * bound thread on Capability 0 pretty soon, so we don't want a
2597 * worker task hogging it.
2598 */
2599 {
2600 nat i;
2601 Capability *cap;
2602 for (i = 1; i < n_capabilities; i++) {
2603 cap = &capabilities[i];
2604 ACQUIRE_LOCK(&cap->lock);
2605 startWorkerTask(cap, workerStart);
2606 RELEASE_LOCK(&cap->lock);
2607 }
2608 }
2609 #endif
2610
2611 trace(TRACE_sched, "start: %d capabilities", n_capabilities);
2612
2613 RELEASE_LOCK(&sched_mutex);
2614 }
2615
2616 void
2617 exitScheduler(
2618 rtsBool wait_foreign
2619 #if !defined(THREADED_RTS)
2620 __attribute__((unused))
2621 #endif
2622 )
2623 /* see Capability.c, shutdownCapability() */
2624 {
2625 Task *task = NULL;
2626
2627 #if defined(THREADED_RTS)
2628 ACQUIRE_LOCK(&sched_mutex);
2629 task = newBoundTask();
2630 RELEASE_LOCK(&sched_mutex);
2631 #endif
2632
2633 // If we haven't killed all the threads yet, do it now.
2634 if (sched_state < SCHED_SHUTTING_DOWN) {
2635 sched_state = SCHED_INTERRUPTING;
2636 scheduleDoGC(NULL,task,rtsFalse);
2637 }
2638 sched_state = SCHED_SHUTTING_DOWN;
2639
2640 #if defined(THREADED_RTS)
2641 {
2642 nat i;
2643
2644 for (i = 0; i < n_capabilities; i++) {
2645 shutdownCapability(&capabilities[i], task, wait_foreign);
2646 }
2647 boundTaskExiting(task);
2648 stopTaskManager();
2649 }
2650 #else
2651 freeCapability(&MainCapability);
2652 #endif
2653 }
2654
2655 void
2656 freeScheduler( void )
2657 {
2658 freeTaskManager();
2659 if (n_capabilities != 1) {
2660 stgFree(capabilities);
2661 }
2662 #if defined(THREADED_RTS)
2663 closeMutex(&sched_mutex);
2664 #endif
2665 }
2666
2667 /* ---------------------------------------------------------------------------
2668 Where are the roots that we know about?
2669
2670 - all the threads on the runnable queue
2671 - all the threads on the blocked queue
2672 - all the threads on the sleeping queue
2673 - all the thread currently executing a _ccall_GC
2674 - all the "main threads"
2675
2676 ------------------------------------------------------------------------ */
2677
2678 /* This has to be protected either by the scheduler monitor, or by the
2679 garbage collection monitor (probably the latter).
2680 KH @ 25/10/99
2681 */
2682
2683 void
2684 GetRoots( evac_fn evac )
2685 {
2686 nat i;
2687 Capability *cap;
2688 Task *task;
2689
2690 #if defined(GRAN)
2691 for (i=0; i<=RtsFlags.GranFlags.proc; i++) {
2692 if ((run_queue_hds[i] != END_TSO_QUEUE) && ((run_queue_hds[i] != NULL)))
2693 evac((StgClosure **)&run_queue_hds[i]);
2694 if ((run_queue_tls[i] != END_TSO_QUEUE) && ((run_queue_tls[i] != NULL)))
2695 evac((StgClosure **)&run_queue_tls[i]);
2696
2697 if ((blocked_queue_hds[i] != END_TSO_QUEUE) && ((blocked_queue_hds[i] != NULL)))
2698 evac((StgClosure **)&blocked_queue_hds[i]);
2699 if ((blocked_queue_tls[i] != END_TSO_QUEUE) && ((blocked_queue_tls[i] != NULL)))
2700 evac((StgClosure **)&blocked_queue_tls[i]);
2701 if ((ccalling_threadss[i] != END_TSO_QUEUE) && ((ccalling_threadss[i] != NULL)))
2702 evac((StgClosure **)&ccalling_threads[i]);
2703 }
2704
2705 markEventQueue();
2706
2707 #else /* !GRAN */
2708
2709 for (i = 0; i < n_capabilities; i++) {
2710 cap = &capabilities[i];
2711 evac((StgClosure **)(void *)&cap->run_queue_hd);
2712 evac((StgClosure **)(void *)&cap->run_queue_tl);
2713 #if defined(THREADED_RTS)
2714 evac((StgClosure **)(void *)&cap->wakeup_queue_hd);
2715 evac((StgClosure **)(void *)&cap->wakeup_queue_tl);
2716 #endif
2717 for (task = cap->suspended_ccalling_tasks; task != NULL;
2718 task=task->next) {
2719 debugTrace(DEBUG_sched,
2720 "evac'ing suspended TSO %lu", (unsigned long)task->suspended_tso->id);
2721 evac((StgClosure **)(void *)&task->suspended_tso);
2722 }
2723
2724 }
2725
2726
2727 #if !defined(THREADED_RTS)
2728 evac((StgClosure **)(void *)&blocked_queue_hd);
2729 evac((StgClosure **)(void *)&blocked_queue_tl);
2730 evac((StgClosure **)(void *)&sleeping_queue);
2731 #endif
2732 #endif
2733
2734 // evac((StgClosure **)&blackhole_queue);
2735
2736 #if defined(THREADED_RTS) || defined(PARALLEL_HASKELL) || defined(GRAN)
2737 markSparkQueue(evac);
2738 #endif
2739
2740 #if defined(RTS_USER_SIGNALS)
2741 // mark the signal handlers (signals should be already blocked)
2742 if (RtsFlags.MiscFlags.install_signal_handlers) {
2743 markSignalHandlers(evac);
2744 }
2745 #endif
2746 }
2747
2748 /* -----------------------------------------------------------------------------
2749 performGC
2750
2751 This is the interface to the garbage collector from Haskell land.
2752 We provide this so that external C code can allocate and garbage
2753 collect when called from Haskell via _ccall_GC.
2754 -------------------------------------------------------------------------- */
2755
2756 static void
2757 performGC_(rtsBool force_major)
2758 {
2759 Task *task;
2760 // We must grab a new Task here, because the existing Task may be
2761 // associated with a particular Capability, and chained onto the
2762 // suspended_ccalling_tasks queue.
2763 ACQUIRE_LOCK(&sched_mutex);
2764 task = newBoundTask();
2765 RELEASE_LOCK(&sched_mutex);
2766 scheduleDoGC(NULL,task,force_major);
2767 boundTaskExiting(task);
2768 }
2769
2770 void
2771 performGC(void)
2772 {
2773 performGC_(rtsFalse);
2774 }
2775
2776 void
2777 performMajorGC(void)
2778 {
2779 performGC_(rtsTrue);
2780 }
2781
2782 /* -----------------------------------------------------------------------------
2783 Stack overflow
2784
2785 If the thread has reached its maximum stack size, then raise the
2786 StackOverflow exception in the offending thread. Otherwise
2787 relocate the TSO into a larger chunk of memory and adjust its stack
2788 size appropriately.
2789 -------------------------------------------------------------------------- */
2790
2791 static StgTSO *
2792 threadStackOverflow(Capability *cap, StgTSO *tso)
2793 {
2794 nat new_stack_size, stack_words;
2795 lnat new_tso_size;
2796 StgPtr new_sp;
2797 StgTSO *dest;
2798
2799 IF_DEBUG(sanity,checkTSO(tso));
2800
2801 // don't allow throwTo() to modify the blocked_exceptions queue
2802 // while we are moving the TSO:
2803 lockClosure((StgClosure *)tso);
2804
2805 if (tso->stack_size >= tso->max_stack_size && !(tso->flags & TSO_BLOCKEX)) {
2806 // NB. never raise a StackOverflow exception if the thread is
2807 // inside Control.Exceptino.block. It is impractical to protect
2808 // against stack overflow exceptions, since virtually anything
2809 // can raise one (even 'catch'), so this is the only sensible
2810 // thing to do here. See bug #767.
2811
2812 debugTrace(DEBUG_gc,
2813 "threadStackOverflow of TSO %ld (%p): stack too large (now %ld; max is %ld)",
2814 (long)tso->id, tso, (long)tso->stack_size, (long)tso->max_stack_size);
2815 IF_DEBUG(gc,
2816 /* If we're debugging, just print out the top of the stack */
2817 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2818 tso->sp+64)));
2819
2820 // Send this thread the StackOverflow exception
2821 unlockTSO(tso);
2822 throwToSingleThreaded(cap, tso, (StgClosure *)stackOverflow_closure);
2823 return tso;
2824 }
2825
2826 /* Try to double the current stack size. If that takes us over the
2827 * maximum stack size for this thread, then use the maximum instead.
2828 * Finally round up so the TSO ends up as a whole number of blocks.
2829 */
2830 new_stack_size = stg_min(tso->stack_size * 2, tso->max_stack_size);
2831 new_tso_size = (lnat)BLOCK_ROUND_UP(new_stack_size * sizeof(W_) +
2832 TSO_STRUCT_SIZE)/sizeof(W_);
2833 new_tso_size = round_to_mblocks(new_tso_size); /* Be MBLOCK-friendly */
2834 new_stack_size = new_tso_size - TSO_STRUCT_SIZEW;
2835
2836 debugTrace(DEBUG_sched,
2837 "increasing stack size from %ld words to %d.",
2838 (long)tso->stack_size, new_stack_size);
2839
2840 dest = (StgTSO *)allocateLocal(cap,new_tso_size);
2841 TICK_ALLOC_TSO(new_stack_size,0);
2842
2843 /* copy the TSO block and the old stack into the new area */
2844 memcpy(dest,tso,TSO_STRUCT_SIZE);
2845 stack_words = tso->stack + tso->stack_size - tso->sp;
2846 new_sp = (P_)dest + new_tso_size - stack_words;
2847 memcpy(new_sp, tso->sp, stack_words * sizeof(W_));
2848
2849 /* relocate the stack pointers... */
2850 dest->sp = new_sp;
2851 dest->stack_size = new_stack_size;
2852
2853 /* Mark the old TSO as relocated. We have to check for relocated
2854 * TSOs in the garbage collector and any primops that deal with TSOs.
2855 *
2856 * It's important to set the sp value to just beyond the end
2857 * of the stack, so we don't attempt to scavenge any part of the
2858 * dead TSO's stack.
2859 */
2860 tso->what_next = ThreadRelocated;
2861 tso->link = dest;
2862 tso->sp = (P_)&(tso->stack[tso->stack_size]);
2863 tso->why_blocked = NotBlocked;
2864
2865 IF_PAR_DEBUG(verbose,
2866 debugBelch("@@ threadStackOverflow of TSO %d (now at %p): stack size increased to %ld\n",
2867 tso->id, tso, tso->stack_size);
2868 /* If we're debugging, just print out the top of the stack */
2869 printStackChunk(tso->sp, stg_min(tso->stack+tso->stack_size,
2870 tso->sp+64)));
2871
2872 unlockTSO(dest);
2873 unlockTSO(tso);
2874
2875 IF_DEBUG(sanity,checkTSO(dest));
2876 #if 0
2877 IF_DEBUG(scheduler,printTSO(dest));
2878 #endif
2879
2880 return dest;
2881 }
2882
2883 /* ---------------------------------------------------------------------------
2884 Interrupt execution
2885 - usually called inside a signal handler so it mustn't do anything fancy.
2886 ------------------------------------------------------------------------ */
2887
2888 void
2889 interruptStgRts(void)
2890 {
2891 sched_state = SCHED_INTERRUPTING;
2892 context_switch = 1;
2893 wakeUpRts();
2894 }
2895
2896 /* -----------------------------------------------------------------------------
2897 Wake up the RTS
2898
2899 This function causes at least one OS thread to wake up and run the
2900 scheduler loop. It is invoked when the RTS might be deadlocked, or
2901 an external event has arrived that may need servicing (eg. a
2902 keyboard interrupt).
2903
2904 In the single-threaded RTS we don't do anything here; we only have
2905 one thread anyway, and the event that caused us to want to wake up
2906 will have interrupted any blocking system call in progress anyway.
2907 -------------------------------------------------------------------------- */
2908
2909 void
2910 wakeUpRts(void)
2911 {
2912 #if defined(THREADED_RTS)
2913 // This forces the IO Manager thread to wakeup, which will
2914 // in turn ensure that some OS thread wakes up and runs the
2915 // scheduler loop, which will cause a GC and deadlock check.
2916 ioManagerWakeup();
2917 #endif
2918 }
2919
2920 /* -----------------------------------------------------------------------------
2921 * checkBlackHoles()
2922 *
2923 * Check the blackhole_queue for threads that can be woken up. We do
2924 * this periodically: before every GC, and whenever the run queue is
2925 * empty.
2926 *
2927 * An elegant solution might be to just wake up all the blocked
2928 * threads with awakenBlockedQueue occasionally: they'll go back to
2929 * sleep again if the object is still a BLACKHOLE. Unfortunately this
2930 * doesn't give us a way to tell whether we've actually managed to
2931 * wake up any threads, so we would be busy-waiting.
2932 *
2933 * -------------------------------------------------------------------------- */
2934
2935 static rtsBool
2936 checkBlackHoles (Capability *cap)
2937 {
2938 StgTSO **prev, *t;
2939 rtsBool any_woke_up = rtsFalse;
2940 StgHalfWord type;
2941
2942 // blackhole_queue is global:
2943 ASSERT_LOCK_HELD(&sched_mutex);
2944
2945 debugTrace(DEBUG_sched, "checking threads blocked on black holes");
2946
2947 // ASSUMES: sched_mutex
2948 prev = &blackhole_queue;
2949 t = blackhole_queue;
2950 while (t != END_TSO_QUEUE) {
2951 ASSERT(t->why_blocked == BlockedOnBlackHole);
2952 type = get_itbl(t->block_info.closure)->type;
2953 if (type != BLACKHOLE && type != CAF_BLACKHOLE) {
2954 IF_DEBUG(sanity,checkTSO(t));
2955 t = unblockOne(cap, t);
2956 // urk, the threads migrate to the current capability
2957 // here, but we'd like to keep them on the original one.
2958 *prev = t;
2959 any_woke_up = rtsTrue;
2960 } else {
2961 prev = &t->link;
2962 t = t->link;
2963 }
2964 }
2965
2966 return any_woke_up;
2967 }
2968
2969 /* -----------------------------------------------------------------------------
2970 Deleting threads
2971
2972 This is used for interruption (^C) and forking, and corresponds to
2973 raising an exception but without letting the thread catch the
2974 exception.
2975 -------------------------------------------------------------------------- */
2976
2977 static void
2978 deleteThread (Capability *cap, StgTSO *tso)
2979 {
2980 // NOTE: must only be called on a TSO that we have exclusive
2981 // access to, because we will call throwToSingleThreaded() below.
2982 // The TSO must be on the run queue of the Capability we own, or
2983 // we must own all Capabilities.
2984
2985 if (tso->why_blocked != BlockedOnCCall &&
2986 tso->why_blocked != BlockedOnCCall_NoUnblockExc) {
2987 throwToSingleThreaded(cap,tso,NULL);
2988 }
2989 }
2990
2991 #ifdef FORKPROCESS_PRIMOP_SUPPORTED
2992 static void
2993 deleteThread_(Capability *cap, StgTSO *tso)
2994 { // for forkProcess only:
2995 // like deleteThread(), but we delete threads in foreign calls, too.
2996
2997 if (tso->why_blocked == BlockedOnCCall ||
2998 tso->why_blocked == BlockedOnCCall_NoUnblockExc) {
2999 unblockOne(cap,tso);
3000 tso->what_next = ThreadKilled;
3001 } else {
3002 deleteThread(cap,tso);
3003 }
3004 }
3005 #endif
3006
3007 /* -----------------------------------------------------------------------------
3008 raiseExceptionHelper
3009
3010 This function is called by the raise# primitve, just so that we can
3011 move some of the tricky bits of raising an exception from C-- into
3012 C. Who knows, it might be a useful re-useable thing here too.
3013 -------------------------------------------------------------------------- */
3014
3015 StgWord
3016 raiseExceptionHelper (StgRegTable *reg, StgTSO *tso, StgClosure *exception)
3017 {
3018 Capability *cap = regTableToCapability(reg);
3019 StgThunk *raise_closure = NULL;
3020 StgPtr p, next;
3021 StgRetInfoTable *info;
3022 //
3023 // This closure represents the expression 'raise# E' where E
3024 // is the exception raise. It is used to overwrite all the
3025 // thunks which are currently under evaluataion.
3026 //
3027
3028 // OLD COMMENT (we don't have MIN_UPD_SIZE now):
3029 // LDV profiling: stg_raise_info has THUNK as its closure
3030 // type. Since a THUNK takes at least MIN_UPD_SIZE words in its
3031 // payload, MIN_UPD_SIZE is more approprate than 1. It seems that
3032 // 1 does not cause any problem unless profiling is performed.
3033 // However, when LDV profiling goes on, we need to linearly scan
3034 // small object pool, where raise_closure is stored, so we should
3035 // use MIN_UPD_SIZE.
3036 //
3037 // raise_closure = (StgClosure *)RET_STGCALL1(P_,allocate,
3038 // sizeofW(StgClosure)+1);
3039 //
3040
3041 //
3042 // Walk up the stack, looking for the catch frame. On the way,
3043 // we update any closures pointed to from update frames with the
3044 // raise closure that we just built.
3045 //
3046 p = tso->sp;
3047 while(1) {
3048 info = get_ret_itbl((StgClosure *)p);
3049 next = p + stack_frame_sizeW((StgClosure *)p);
3050 switch (info->i.type) {
3051
3052 case UPDATE_FRAME:
3053 // Only create raise_closure if we need to.
3054 if (raise_closure == NULL) {
3055 raise_closure =
3056 (StgThunk *)allocateLocal(cap,sizeofW(StgThunk)+1);
3057 SET_HDR(raise_closure, &stg_raise_info, CCCS);
3058 raise_closure->payload[0] = exception;
3059 }
3060 UPD_IND(((StgUpdateFrame *)p)->updatee,(StgClosure *)raise_closure);
3061 p = next;
3062 continue;
3063
3064 case ATOMICALLY_FRAME:
3065 debugTrace(DEBUG_stm, "found ATOMICALLY_FRAME at %p", p);
3066 tso->sp = p;
3067 return ATOMICALLY_FRAME;
3068
3069 case CATCH_FRAME:
3070 tso->sp = p;
3071 return CATCH_FRAME;
3072
3073 case CATCH_STM_FRAME:
3074 debugTrace(DEBUG_stm, "found CATCH_STM_FRAME at %p", p);
3075 tso->sp = p;
3076 return CATCH_STM_FRAME;
3077
3078 case STOP_FRAME:
3079 tso->sp = p;
3080 return STOP_FRAME;
3081
3082 case CATCH_RETRY_FRAME:
3083 default:
3084 p = next;
3085 continue;
3086 }
3087 }
3088 }
3089
3090
3091 /* -----------------------------------------------------------------------------
3092 findRetryFrameHelper
3093
3094 This function is called by the retry# primitive. It traverses the stack
3095 leaving tso->sp referring to the frame which should handle the retry.
3096
3097 This should either be a CATCH_RETRY_FRAME (if the retry# is within an orElse#)
3098 or should be a ATOMICALLY_FRAME (if the retry# reaches the top level).
3099
3100 We skip CATCH_STM_FRAMEs (aborting and rolling back the nested tx that they
3101 create) because retries are not considered to be exceptions, despite the
3102 similar implementation.
3103
3104 We should not expect to see CATCH_FRAME or STOP_FRAME because those should
3105 not be created within memory transactions.
3106 -------------------------------------------------------------------------- */
3107
3108 StgWord
3109 findRetryFrameHelper (StgTSO *tso)
3110 {
3111 StgPtr p, next;
3112 StgRetInfoTable *info;
3113
3114 p = tso -> sp;
3115 while (1) {
3116 info = get_ret_itbl((StgClosure *)p);
3117 next = p + stack_frame_sizeW((StgClosure *)p);
3118 switch (info->i.type) {
3119
3120 case ATOMICALLY_FRAME:
3121 debugTrace(DEBUG_stm,
3122 "found ATOMICALLY_FRAME at %p during retry", p);
3123 tso->sp = p;
3124 return ATOMICALLY_FRAME;
3125
3126 case CATCH_RETRY_FRAME:
3127 debugTrace(DEBUG_stm,
3128 "found CATCH_RETRY_FRAME at %p during retrry", p);
3129 tso->sp = p;
3130 return CATCH_RETRY_FRAME;
3131
3132 case CATCH_STM_FRAME: {
3133 StgTRecHeader *trec = tso -> trec;
3134 StgTRecHeader *outer = stmGetEnclosingTRec(trec);
3135 debugTrace(DEBUG_stm,
3136 "found CATCH_STM_FRAME at %p during retry", p);
3137 debugTrace(DEBUG_stm, "trec=%p outer=%p", trec, outer);
3138 stmAbortTransaction(tso -> cap, trec);
3139 stmFreeAbortedTRec(tso -> cap, trec);
3140 tso -> trec = outer;
3141 p = next;
3142 continue;
3143 }
3144
3145
3146 default:
3147 ASSERT(info->i.type != CATCH_FRAME);
3148 ASSERT(info->i.type != STOP_FRAME);
3149 p = next;
3150 continue;
3151 }
3152 }
3153 }
3154
3155 /* -----------------------------------------------------------------------------
3156 resurrectThreads is called after garbage collection on the list of
3157 threads found to be garbage. Each of these threads will be woken
3158 up and sent a signal: BlockedOnDeadMVar if the thread was blocked
3159 on an MVar, or NonTermination if the thread was blocked on a Black
3160 Hole.
3161
3162 Locks: assumes we hold *all* the capabilities.
3163 -------------------------------------------------------------------------- */
3164
3165 void
3166 resurrectThreads (StgTSO *threads)
3167 {
3168 StgTSO *tso, *next;
3169 Capability *cap;
3170
3171 for (tso = threads; tso != END_TSO_QUEUE; tso = next) {
3172 next = tso->global_link;
3173 tso->global_link = all_threads;
3174 all_threads = tso;
3175 debugTrace(DEBUG_sched, "resurrecting thread %lu", (unsigned long)tso->id);
3176
3177 // Wake up the thread on the Capability it was last on
3178 cap = tso->cap;
3179
3180 switch (tso->why_blocked) {
3181 case BlockedOnMVar:
3182 case BlockedOnException:
3183 /* Called by GC - sched_mutex lock is currently held. */
3184 throwToSingleThreaded(cap, tso,
3185 (StgClosure *)BlockedOnDeadMVar_closure);
3186 break;
3187 case BlockedOnBlackHole:
3188 throwToSingleThreaded(cap, tso,
3189 (StgClosure *)NonTermination_closure);
3190 break;
3191 case BlockedOnSTM:
3192 throwToSingleThreaded(cap, tso,
3193 (StgClosure *)BlockedIndefinitely_closure);
3194 break;
3195 case NotBlocked:
3196 /* This might happen if the thread was blocked on a black hole
3197 * belonging to a thread that we've just woken up (raiseAsync
3198 * can wake up threads, remember...).
3199 */
3200 continue;
3201 default:
3202 barf("resurrectThreads: thread blocked in a strange way");
3203 }
3204 }
3205 }