Add events to show when GC threads are idle/working
[ghc.git] / rts / sm / GC.c
1 /* -----------------------------------------------------------------------------
2 *
3 * (c) The GHC Team 1998-2008
4 *
5 * Generational garbage collector
6 *
7 * Documentation on the architecture of the Garbage Collector can be
8 * found in the online commentary:
9 *
10 * http://hackage.haskell.org/trac/ghc/wiki/Commentary/Rts/Storage/GC
11 *
12 * ---------------------------------------------------------------------------*/
13
14 #include "PosixSource.h"
15 #include "Rts.h"
16 #include "HsFFI.h"
17
18 #include "Storage.h"
19 #include "RtsUtils.h"
20 #include "Apply.h"
21 #include "Updates.h"
22 #include "Stats.h"
23 #include "Schedule.h"
24 #include "Sanity.h"
25 #include "BlockAlloc.h"
26 #include "ProfHeap.h"
27 #include "Weak.h"
28 #include "Prelude.h"
29 #include "RtsSignals.h"
30 #include "STM.h"
31 #if defined(RTS_GTK_FRONTPANEL)
32 #include "FrontPanel.h"
33 #endif
34 #include "Trace.h"
35 #include "RetainerProfile.h"
36 #include "LdvProfile.h"
37 #include "RaiseAsync.h"
38 #include "Papi.h"
39 #include "Stable.h"
40
41 #include "GC.h"
42 #include "GCThread.h"
43 #include "Compact.h"
44 #include "Evac.h"
45 #include "Scav.h"
46 #include "GCUtils.h"
47 #include "MarkStack.h"
48 #include "MarkWeak.h"
49 #include "Sparks.h"
50 #include "Sweep.h"
51
52 #include <string.h> // for memset()
53 #include <unistd.h>
54
55 /* -----------------------------------------------------------------------------
56 Global variables
57 -------------------------------------------------------------------------- */
58
59 /* STATIC OBJECT LIST.
60 *
61 * During GC:
62 * We maintain a linked list of static objects that are still live.
63 * The requirements for this list are:
64 *
65 * - we need to scan the list while adding to it, in order to
66 * scavenge all the static objects (in the same way that
67 * breadth-first scavenging works for dynamic objects).
68 *
69 * - we need to be able to tell whether an object is already on
70 * the list, to break loops.
71 *
72 * Each static object has a "static link field", which we use for
73 * linking objects on to the list. We use a stack-type list, consing
74 * objects on the front as they are added (this means that the
75 * scavenge phase is depth-first, not breadth-first, but that
76 * shouldn't matter).
77 *
78 * A separate list is kept for objects that have been scavenged
79 * already - this is so that we can zero all the marks afterwards.
80 *
81 * An object is on the list if its static link field is non-zero; this
82 * means that we have to mark the end of the list with '1', not NULL.
83 *
84 * Extra notes for generational GC:
85 *
86 * Each generation has a static object list associated with it. When
87 * collecting generations up to N, we treat the static object lists
88 * from generations > N as roots.
89 *
90 * We build up a static object list while collecting generations 0..N,
91 * which is then appended to the static object list of generation N+1.
92 */
93
94 /* N is the oldest generation being collected, where the generations
95 * are numbered starting at 0. A major GC (indicated by the major_gc
96 * flag) is when we're collecting all generations. We only attempt to
97 * deal with static objects and GC CAFs when doing a major GC.
98 */
99 nat N;
100 rtsBool major_gc;
101
102 /* Data used for allocation area sizing.
103 */
104 static lnat g0s0_pcnt_kept = 30; // percentage of g0s0 live at last minor GC
105
106 /* Mut-list stats */
107 #ifdef DEBUG
108 nat mutlist_MUTVARS,
109 mutlist_MUTARRS,
110 mutlist_MVARS,
111 mutlist_OTHERS;
112 #endif
113
114 /* Thread-local data for each GC thread
115 */
116 gc_thread **gc_threads = NULL;
117
118 #if !defined(THREADED_RTS)
119 StgWord8 the_gc_thread[sizeof(gc_thread) + 64 * sizeof(step_workspace)];
120 #endif
121
122 // Number of threads running in *this* GC. Affects how many
123 // step->todos[] lists we have to look in to find work.
124 nat n_gc_threads;
125
126 // For stats:
127 long copied; // *words* copied & scavenged during this GC
128
129 rtsBool work_stealing;
130
131 DECLARE_GCT
132
133 /* -----------------------------------------------------------------------------
134 Static function declarations
135 -------------------------------------------------------------------------- */
136
137 static void mark_root (void *user, StgClosure **root);
138 static void zero_static_object_list (StgClosure* first_static);
139 static nat initialise_N (rtsBool force_major_gc);
140 static void init_collected_gen (nat g, nat threads);
141 static void init_uncollected_gen (nat g, nat threads);
142 static void init_gc_thread (gc_thread *t);
143 static void update_task_list (void);
144 static void resize_generations (void);
145 static void resize_nursery (void);
146 static void start_gc_threads (void);
147 static void scavenge_until_all_done (void);
148 static StgWord inc_running (void);
149 static StgWord dec_running (void);
150 static void wakeup_gc_threads (nat n_threads, nat me);
151 static void shutdown_gc_threads (nat n_threads, nat me);
152
153 #if 0 && defined(DEBUG)
154 static void gcCAFs (void);
155 #endif
156
157 /* -----------------------------------------------------------------------------
158 The mark stack.
159 -------------------------------------------------------------------------- */
160
161 bdescr *mark_stack_top_bd; // topmost block in the mark stack
162 bdescr *mark_stack_bd; // current block in the mark stack
163 StgPtr mark_sp; // pointer to the next unallocated mark stack entry
164
165 /* -----------------------------------------------------------------------------
166 GarbageCollect: the main entry point to the garbage collector.
167
168 Locks held: all capabilities are held throughout GarbageCollect().
169 -------------------------------------------------------------------------- */
170
171 void
172 GarbageCollect (rtsBool force_major_gc,
173 nat gc_type USED_IF_THREADS,
174 Capability *cap)
175 {
176 bdescr *bd;
177 step *stp;
178 lnat live, allocated, max_copied, avg_copied, slop;
179 gc_thread *saved_gct;
180 nat g, s, t, n;
181
182 // necessary if we stole a callee-saves register for gct:
183 saved_gct = gct;
184
185 #ifdef PROFILING
186 CostCentreStack *prev_CCS;
187 #endif
188
189 ACQUIRE_SM_LOCK;
190
191 #if defined(RTS_USER_SIGNALS)
192 if (RtsFlags.MiscFlags.install_signal_handlers) {
193 // block signals
194 blockUserSignals();
195 }
196 #endif
197
198 ASSERT(sizeof(step_workspace) == 16 * sizeof(StgWord));
199 // otherwise adjust the padding in step_workspace.
200
201 // tell the stats department that we've started a GC
202 stat_startGC();
203
204 // tell the STM to discard any cached closures it's hoping to re-use
205 stmPreGCHook();
206
207 // lock the StablePtr table
208 stablePtrPreGC();
209
210 #ifdef DEBUG
211 mutlist_MUTVARS = 0;
212 mutlist_MUTARRS = 0;
213 mutlist_OTHERS = 0;
214 #endif
215
216 // attribute any costs to CCS_GC
217 #ifdef PROFILING
218 prev_CCS = CCCS;
219 CCCS = CCS_GC;
220 #endif
221
222 /* Approximate how much we allocated.
223 * Todo: only when generating stats?
224 */
225 allocated = calcAllocated();
226
227 /* Figure out which generation to collect
228 */
229 n = initialise_N(force_major_gc);
230
231 #if defined(THREADED_RTS)
232 work_stealing = RtsFlags.ParFlags.parGcLoadBalancingEnabled &&
233 N >= RtsFlags.ParFlags.parGcLoadBalancingGen;
234 // It's not always a good idea to do load balancing in parallel
235 // GC. In particular, for a parallel program we don't want to
236 // lose locality by moving cached data into another CPU's cache
237 // (this effect can be quite significant).
238 //
239 // We could have a more complex way to deterimine whether to do
240 // work stealing or not, e.g. it might be a good idea to do it
241 // if the heap is big. For now, we just turn it on or off with
242 // a flag.
243 #endif
244
245 /* Start threads, so they can be spinning up while we finish initialisation.
246 */
247 start_gc_threads();
248
249 #if defined(THREADED_RTS)
250 /* How many threads will be participating in this GC?
251 * We don't try to parallelise minor GCs (unless the user asks for
252 * it with +RTS -gn0), or mark/compact/sweep GC.
253 */
254 if (gc_type == PENDING_GC_PAR) {
255 n_gc_threads = RtsFlags.ParFlags.nNodes;
256 } else {
257 n_gc_threads = 1;
258 }
259 #else
260 n_gc_threads = 1;
261 #endif
262
263 debugTrace(DEBUG_gc, "GC (gen %d): %d KB to collect, %ld MB in use, using %d thread(s)",
264 N, n * (BLOCK_SIZE / 1024), mblocks_allocated, n_gc_threads);
265
266 #ifdef RTS_GTK_FRONTPANEL
267 if (RtsFlags.GcFlags.frontpanel) {
268 updateFrontPanelBeforeGC(N);
269 }
270 #endif
271
272 #ifdef DEBUG
273 // check for memory leaks if DEBUG is on
274 memInventory(DEBUG_gc);
275 #endif
276
277 // check stack sanity *before* GC
278 IF_DEBUG(sanity, checkFreeListSanity());
279 IF_DEBUG(sanity, checkMutableLists(rtsTrue));
280
281 // Initialise all our gc_thread structures
282 for (t = 0; t < n_gc_threads; t++) {
283 init_gc_thread(gc_threads[t]);
284 }
285
286 // Initialise all the generations/steps that we're collecting.
287 for (g = 0; g <= N; g++) {
288 init_collected_gen(g,n_gc_threads);
289 }
290
291 // Initialise all the generations/steps that we're *not* collecting.
292 for (g = N+1; g < RtsFlags.GcFlags.generations; g++) {
293 init_uncollected_gen(g,n_gc_threads);
294 }
295
296 /* Allocate a mark stack if we're doing a major collection.
297 */
298 if (major_gc && oldest_gen->steps[0].mark) {
299 mark_stack_bd = allocBlock();
300 mark_stack_top_bd = mark_stack_bd;
301 mark_stack_bd->link = NULL;
302 mark_stack_bd->u.back = NULL;
303 mark_sp = mark_stack_bd->start;
304 } else {
305 mark_stack_bd = NULL;
306 mark_stack_top_bd = NULL;
307 mark_sp = NULL;
308 }
309
310 // this is the main thread
311 #ifdef THREADED_RTS
312 if (n_gc_threads == 1) {
313 SET_GCT(gc_threads[0]);
314 } else {
315 SET_GCT(gc_threads[cap->no]);
316 }
317 #else
318 SET_GCT(gc_threads[0]);
319 #endif
320
321 /* -----------------------------------------------------------------------
322 * follow all the roots that we know about:
323 */
324
325 // the main thread is running: this prevents any other threads from
326 // exiting prematurely, so we can start them now.
327 // NB. do this after the mutable lists have been saved above, otherwise
328 // the other GC threads will be writing into the old mutable lists.
329 inc_running();
330 wakeup_gc_threads(n_gc_threads, gct->thread_index);
331
332 // Mutable lists from each generation > N
333 // we want to *scavenge* these roots, not evacuate them: they're not
334 // going to move in this GC.
335 // Also do them in reverse generation order, for the usual reason:
336 // namely to reduce the likelihood of spurious old->new pointers.
337 //
338 for (g = RtsFlags.GcFlags.generations-1; g > N; g--) {
339 scavenge_mutable_list(generations[g].saved_mut_list, &generations[g]);
340 freeChain_sync(generations[g].saved_mut_list);
341 generations[g].saved_mut_list = NULL;
342
343 }
344
345 // scavenge the capability-private mutable lists. This isn't part
346 // of markSomeCapabilities() because markSomeCapabilities() can only
347 // call back into the GC via mark_root() (due to the gct register
348 // variable).
349 if (n_gc_threads == 1) {
350 for (n = 0; n < n_capabilities; n++) {
351 scavenge_capability_mut_lists(&capabilities[n]);
352 }
353 } else {
354 scavenge_capability_mut_lists(&capabilities[gct->thread_index]);
355 }
356
357 // follow roots from the CAF list (used by GHCi)
358 gct->evac_step = 0;
359 markCAFs(mark_root, gct);
360
361 // follow all the roots that the application knows about.
362 gct->evac_step = 0;
363 markSomeCapabilities(mark_root, gct, gct->thread_index, n_gc_threads,
364 rtsTrue/*prune sparks*/);
365
366 #if defined(RTS_USER_SIGNALS)
367 // mark the signal handlers (signals should be already blocked)
368 markSignalHandlers(mark_root, gct);
369 #endif
370
371 // Mark the weak pointer list, and prepare to detect dead weak pointers.
372 markWeakPtrList();
373 initWeakForGC();
374
375 // Mark the stable pointer table.
376 markStablePtrTable(mark_root, gct);
377
378 /* -------------------------------------------------------------------------
379 * Repeatedly scavenge all the areas we know about until there's no
380 * more scavenging to be done.
381 */
382 for (;;)
383 {
384 scavenge_until_all_done();
385 // The other threads are now stopped. We might recurse back to
386 // here, but from now on this is the only thread.
387
388 // if any blackholes are alive, make the threads that wait on
389 // them alive too.
390 if (traverseBlackholeQueue()) {
391 inc_running();
392 continue;
393 }
394
395 // must be last... invariant is that everything is fully
396 // scavenged at this point.
397 if (traverseWeakPtrList()) { // returns rtsTrue if evaced something
398 inc_running();
399 continue;
400 }
401
402 // If we get to here, there's really nothing left to do.
403 break;
404 }
405
406 shutdown_gc_threads(n_gc_threads, gct->thread_index);
407
408 // Update pointers from the Task list
409 update_task_list();
410
411 // Now see which stable names are still alive.
412 gcStablePtrTable();
413
414 #ifdef PROFILING
415 // We call processHeapClosureForDead() on every closure destroyed during
416 // the current garbage collection, so we invoke LdvCensusForDead().
417 if (RtsFlags.ProfFlags.doHeapProfile == HEAP_BY_LDV
418 || RtsFlags.ProfFlags.bioSelector != NULL)
419 LdvCensusForDead(N);
420 #endif
421
422 // NO MORE EVACUATION AFTER THIS POINT!
423
424 // Two-space collector: free the old to-space.
425 // g0s0->old_blocks is the old nursery
426 // g0s0->blocks is to-space from the previous GC
427 if (RtsFlags.GcFlags.generations == 1) {
428 if (g0s0->blocks != NULL) {
429 freeChain(g0s0->blocks);
430 g0s0->blocks = NULL;
431 }
432 }
433
434 // For each workspace, in each thread, move the copied blocks to the step
435 {
436 gc_thread *thr;
437 step_workspace *ws;
438 bdescr *prev, *next;
439
440 for (t = 0; t < n_gc_threads; t++) {
441 thr = gc_threads[t];
442
443 // not step 0
444 if (RtsFlags.GcFlags.generations == 1) {
445 s = 0;
446 } else {
447 s = 1;
448 }
449 for (; s < total_steps; s++) {
450 ws = &thr->steps[s];
451
452 // Push the final block
453 if (ws->todo_bd) {
454 push_scanned_block(ws->todo_bd, ws);
455 }
456
457 ASSERT(gct->scan_bd == NULL);
458 ASSERT(countBlocks(ws->scavd_list) == ws->n_scavd_blocks);
459
460 prev = NULL;
461 for (bd = ws->scavd_list; bd != NULL; bd = bd->link) {
462 ws->step->n_words += bd->free - bd->start;
463 prev = bd;
464 }
465 if (prev != NULL) {
466 prev->link = ws->step->blocks;
467 ws->step->blocks = ws->scavd_list;
468 }
469 ws->step->n_blocks += ws->n_scavd_blocks;
470 }
471 }
472
473 // Add all the partial blocks *after* we've added all the full
474 // blocks. This is so that we can grab the partial blocks back
475 // again and try to fill them up in the next GC.
476 for (t = 0; t < n_gc_threads; t++) {
477 thr = gc_threads[t];
478
479 // not step 0
480 if (RtsFlags.GcFlags.generations == 1) {
481 s = 0;
482 } else {
483 s = 1;
484 }
485 for (; s < total_steps; s++) {
486 ws = &thr->steps[s];
487
488 prev = NULL;
489 for (bd = ws->part_list; bd != NULL; bd = next) {
490 next = bd->link;
491 if (bd->free == bd->start) {
492 if (prev == NULL) {
493 ws->part_list = next;
494 } else {
495 prev->link = next;
496 }
497 freeGroup(bd);
498 ws->n_part_blocks--;
499 } else {
500 ws->step->n_words += bd->free - bd->start;
501 prev = bd;
502 }
503 }
504 if (prev != NULL) {
505 prev->link = ws->step->blocks;
506 ws->step->blocks = ws->part_list;
507 }
508 ws->step->n_blocks += ws->n_part_blocks;
509
510 ASSERT(countBlocks(ws->step->blocks) == ws->step->n_blocks);
511 ASSERT(countOccupied(ws->step->blocks) == ws->step->n_words);
512 }
513 }
514 }
515
516 // Finally: compact or sweep the oldest generation.
517 if (major_gc && oldest_gen->steps[0].mark) {
518 if (oldest_gen->steps[0].compact)
519 compact(gct->scavenged_static_objects);
520 else
521 sweep(&oldest_gen->steps[0]);
522 }
523
524 /* run through all the generations/steps and tidy up
525 */
526 copied = 0;
527 max_copied = 0;
528 avg_copied = 0;
529 {
530 nat i;
531 for (i=0; i < n_gc_threads; i++) {
532 if (n_gc_threads > 1) {
533 debugTrace(DEBUG_gc,"thread %d:", i);
534 debugTrace(DEBUG_gc," copied %ld", gc_threads[i]->copied * sizeof(W_));
535 debugTrace(DEBUG_gc," scanned %ld", gc_threads[i]->scanned * sizeof(W_));
536 debugTrace(DEBUG_gc," any_work %ld", gc_threads[i]->any_work);
537 debugTrace(DEBUG_gc," no_work %ld", gc_threads[i]->no_work);
538 debugTrace(DEBUG_gc," scav_find_work %ld", gc_threads[i]->scav_find_work);
539 }
540 copied += gc_threads[i]->copied;
541 max_copied = stg_max(gc_threads[i]->copied, max_copied);
542 }
543 if (n_gc_threads == 1) {
544 max_copied = 0;
545 avg_copied = 0;
546 } else {
547 avg_copied = copied;
548 }
549 }
550
551 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
552
553 if (g == N) {
554 generations[g].collections++; // for stats
555 if (n_gc_threads > 1) generations[g].par_collections++;
556 }
557
558 // Count the mutable list as bytes "copied" for the purposes of
559 // stats. Every mutable list is copied during every GC.
560 if (g > 0) {
561 nat mut_list_size = 0;
562 for (bd = generations[g].mut_list; bd != NULL; bd = bd->link) {
563 mut_list_size += bd->free - bd->start;
564 }
565 for (n = 0; n < n_capabilities; n++) {
566 for (bd = capabilities[n].mut_lists[g];
567 bd != NULL; bd = bd->link) {
568 mut_list_size += bd->free - bd->start;
569 }
570 }
571 copied += mut_list_size;
572
573 debugTrace(DEBUG_gc,
574 "mut_list_size: %lu (%d vars, %d arrays, %d MVARs, %d others)",
575 (unsigned long)(mut_list_size * sizeof(W_)),
576 mutlist_MUTVARS, mutlist_MUTARRS, mutlist_MVARS, mutlist_OTHERS);
577 }
578
579 for (s = 0; s < generations[g].n_steps; s++) {
580 bdescr *next, *prev;
581 stp = &generations[g].steps[s];
582
583 // for generations we collected...
584 if (g <= N) {
585
586 /* free old memory and shift to-space into from-space for all
587 * the collected steps (except the allocation area). These
588 * freed blocks will probaby be quickly recycled.
589 */
590 if (!(g == 0 && s == 0 && RtsFlags.GcFlags.generations > 1)) {
591 if (stp->mark)
592 {
593 // tack the new blocks on the end of the existing blocks
594 if (stp->old_blocks != NULL) {
595
596 prev = NULL;
597 for (bd = stp->old_blocks; bd != NULL; bd = next) {
598
599 next = bd->link;
600
601 if (!(bd->flags & BF_MARKED))
602 {
603 if (prev == NULL) {
604 stp->old_blocks = next;
605 } else {
606 prev->link = next;
607 }
608 freeGroup(bd);
609 stp->n_old_blocks--;
610 }
611 else
612 {
613 stp->n_words += bd->free - bd->start;
614
615 // NB. this step might not be compacted next
616 // time, so reset the BF_MARKED flags.
617 // They are set before GC if we're going to
618 // compact. (search for BF_MARKED above).
619 bd->flags &= ~BF_MARKED;
620
621 // between GCs, all blocks in the heap except
622 // for the nursery have the BF_EVACUATED flag set.
623 bd->flags |= BF_EVACUATED;
624
625 prev = bd;
626 }
627 }
628
629 if (prev != NULL) {
630 prev->link = stp->blocks;
631 stp->blocks = stp->old_blocks;
632 }
633 }
634 // add the new blocks to the block tally
635 stp->n_blocks += stp->n_old_blocks;
636 ASSERT(countBlocks(stp->blocks) == stp->n_blocks);
637 ASSERT(countOccupied(stp->blocks) == stp->n_words);
638 }
639 else // not copacted
640 {
641 freeChain(stp->old_blocks);
642 }
643 stp->old_blocks = NULL;
644 stp->n_old_blocks = 0;
645 }
646
647 /* LARGE OBJECTS. The current live large objects are chained on
648 * scavenged_large, having been moved during garbage
649 * collection from large_objects. Any objects left on
650 * large_objects list are therefore dead, so we free them here.
651 */
652 for (bd = stp->large_objects; bd != NULL; bd = next) {
653 next = bd->link;
654 freeGroup(bd);
655 bd = next;
656 }
657
658 stp->large_objects = stp->scavenged_large_objects;
659 stp->n_large_blocks = stp->n_scavenged_large_blocks;
660
661 }
662 else // for older generations...
663 {
664 /* For older generations, we need to append the
665 * scavenged_large_object list (i.e. large objects that have been
666 * promoted during this GC) to the large_object list for that step.
667 */
668 for (bd = stp->scavenged_large_objects; bd; bd = next) {
669 next = bd->link;
670 dbl_link_onto(bd, &stp->large_objects);
671 }
672
673 // add the new blocks we promoted during this GC
674 stp->n_large_blocks += stp->n_scavenged_large_blocks;
675 }
676 }
677 }
678
679 // update the max size of older generations after a major GC
680 resize_generations();
681
682 // Calculate the amount of live data for stats.
683 live = calcLiveWords();
684
685 // Free the small objects allocated via allocate(), since this will
686 // all have been copied into G0S1 now.
687 if (RtsFlags.GcFlags.generations > 1) {
688 if (g0s0->blocks != NULL) {
689 freeChain(g0s0->blocks);
690 g0s0->blocks = NULL;
691 }
692 g0s0->n_blocks = 0;
693 g0s0->n_words = 0;
694 }
695 alloc_blocks = 0;
696 alloc_blocks_lim = RtsFlags.GcFlags.minAllocAreaSize;
697
698 // Start a new pinned_object_block
699 pinned_object_block = NULL;
700
701 // Free the mark stack.
702 if (mark_stack_top_bd != NULL) {
703 debugTrace(DEBUG_gc, "mark stack: %d blocks",
704 countBlocks(mark_stack_top_bd));
705 freeChain(mark_stack_top_bd);
706 }
707
708 // Free any bitmaps.
709 for (g = 0; g <= N; g++) {
710 for (s = 0; s < generations[g].n_steps; s++) {
711 stp = &generations[g].steps[s];
712 if (stp->bitmap != NULL) {
713 freeGroup(stp->bitmap);
714 stp->bitmap = NULL;
715 }
716 }
717 }
718
719 resize_nursery();
720
721 // mark the garbage collected CAFs as dead
722 #if 0 && defined(DEBUG) // doesn't work at the moment
723 if (major_gc) { gcCAFs(); }
724 #endif
725
726 #ifdef PROFILING
727 // resetStaticObjectForRetainerProfiling() must be called before
728 // zeroing below.
729 if (n_gc_threads > 1) {
730 barf("profiling is currently broken with multi-threaded GC");
731 // ToDo: fix the gct->scavenged_static_objects below
732 }
733 resetStaticObjectForRetainerProfiling(gct->scavenged_static_objects);
734 #endif
735
736 // zero the scavenged static object list
737 if (major_gc) {
738 nat i;
739 for (i = 0; i < n_gc_threads; i++) {
740 zero_static_object_list(gc_threads[i]->scavenged_static_objects);
741 }
742 }
743
744 // Reset the nursery
745 resetNurseries();
746
747 // start any pending finalizers
748 RELEASE_SM_LOCK;
749 scheduleFinalizers(cap, old_weak_ptr_list);
750 ACQUIRE_SM_LOCK;
751
752 // send exceptions to any threads which were about to die
753 RELEASE_SM_LOCK;
754 resurrectThreads(resurrected_threads);
755 performPendingThrowTos(exception_threads);
756 ACQUIRE_SM_LOCK;
757
758 // Update the stable pointer hash table.
759 updateStablePtrTable(major_gc);
760
761 // check sanity after GC
762 IF_DEBUG(sanity, checkSanity());
763
764 // extra GC trace info
765 IF_DEBUG(gc, statDescribeGens());
766
767 #ifdef DEBUG
768 // symbol-table based profiling
769 /* heapCensus(to_blocks); */ /* ToDo */
770 #endif
771
772 // restore enclosing cost centre
773 #ifdef PROFILING
774 CCCS = prev_CCS;
775 #endif
776
777 #ifdef DEBUG
778 // check for memory leaks if DEBUG is on
779 memInventory(DEBUG_gc);
780 #endif
781
782 #ifdef RTS_GTK_FRONTPANEL
783 if (RtsFlags.GcFlags.frontpanel) {
784 updateFrontPanelAfterGC( N, live );
785 }
786 #endif
787
788 // ok, GC over: tell the stats department what happened.
789 slop = calcLiveBlocks() * BLOCK_SIZE_W - live;
790 stat_endGC(allocated, live, copied, N, max_copied, avg_copied, slop);
791
792 // unlock the StablePtr table
793 stablePtrPostGC();
794
795 // Guess which generation we'll collect *next* time
796 initialise_N(force_major_gc);
797
798 #if defined(RTS_USER_SIGNALS)
799 if (RtsFlags.MiscFlags.install_signal_handlers) {
800 // unblock signals again
801 unblockUserSignals();
802 }
803 #endif
804
805 RELEASE_SM_LOCK;
806
807 SET_GCT(saved_gct);
808 }
809
810 /* -----------------------------------------------------------------------------
811 Figure out which generation to collect, initialise N and major_gc.
812
813 Also returns the total number of blocks in generations that will be
814 collected.
815 -------------------------------------------------------------------------- */
816
817 static nat
818 initialise_N (rtsBool force_major_gc)
819 {
820 int g;
821 nat s, blocks, blocks_total;
822
823 blocks = 0;
824 blocks_total = 0;
825
826 if (force_major_gc) {
827 N = RtsFlags.GcFlags.generations - 1;
828 } else {
829 N = 0;
830 }
831
832 for (g = RtsFlags.GcFlags.generations - 1; g >= 0; g--) {
833 blocks = 0;
834 for (s = 0; s < generations[g].n_steps; s++) {
835 blocks += generations[g].steps[s].n_words / BLOCK_SIZE_W;
836 blocks += generations[g].steps[s].n_large_blocks;
837 }
838 if (blocks >= generations[g].max_blocks) {
839 N = stg_max(N,g);
840 }
841 if ((nat)g <= N) {
842 blocks_total += blocks;
843 }
844 }
845
846 blocks_total += countNurseryBlocks();
847
848 major_gc = (N == RtsFlags.GcFlags.generations-1);
849 return blocks_total;
850 }
851
852 /* -----------------------------------------------------------------------------
853 Initialise the gc_thread structures.
854 -------------------------------------------------------------------------- */
855
856 #define GC_THREAD_INACTIVE 0
857 #define GC_THREAD_STANDING_BY 1
858 #define GC_THREAD_RUNNING 2
859 #define GC_THREAD_WAITING_TO_CONTINUE 3
860
861 static void
862 new_gc_thread (nat n, gc_thread *t)
863 {
864 nat s;
865 step_workspace *ws;
866
867 #ifdef THREADED_RTS
868 t->id = 0;
869 initSpinLock(&t->gc_spin);
870 initSpinLock(&t->mut_spin);
871 ACQUIRE_SPIN_LOCK(&t->gc_spin);
872 t->wakeup = GC_THREAD_INACTIVE; // starts true, so we can wait for the
873 // thread to start up, see wakeup_gc_threads
874 #endif
875
876 t->thread_index = n;
877 t->free_blocks = NULL;
878 t->gc_count = 0;
879
880 init_gc_thread(t);
881
882 #ifdef USE_PAPI
883 t->papi_events = -1;
884 #endif
885
886 for (s = 0; s < total_steps; s++)
887 {
888 ws = &t->steps[s];
889 ws->step = &all_steps[s];
890 ASSERT(s == ws->step->abs_no);
891 ws->my_gct = t;
892
893 ws->todo_bd = NULL;
894 ws->todo_q = newWSDeque(128);
895 ws->todo_overflow = NULL;
896 ws->n_todo_overflow = 0;
897
898 ws->part_list = NULL;
899 ws->n_part_blocks = 0;
900
901 ws->scavd_list = NULL;
902 ws->n_scavd_blocks = 0;
903 }
904 }
905
906
907 void
908 initGcThreads (void)
909 {
910 if (gc_threads == NULL) {
911 #if defined(THREADED_RTS)
912 nat i;
913 gc_threads = stgMallocBytes (RtsFlags.ParFlags.nNodes *
914 sizeof(gc_thread*),
915 "alloc_gc_threads");
916
917 for (i = 0; i < RtsFlags.ParFlags.nNodes; i++) {
918 gc_threads[i] =
919 stgMallocBytes(sizeof(gc_thread) + total_steps * sizeof(step_workspace),
920 "alloc_gc_threads");
921
922 new_gc_thread(i, gc_threads[i]);
923 }
924 #else
925 gc_threads = stgMallocBytes (sizeof(gc_thread*),"alloc_gc_threads");
926 gc_threads[0] = gct;
927 new_gc_thread(0,gc_threads[0]);
928 #endif
929 }
930 }
931
932 void
933 freeGcThreads (void)
934 {
935 if (gc_threads != NULL) {
936 #if defined(THREADED_RTS)
937 nat i;
938 for (i = 0; i < RtsFlags.ParFlags.nNodes; i++) {
939 stgFree (gc_threads[i]);
940 }
941 stgFree (gc_threads);
942 #else
943 stgFree (gc_threads);
944 #endif
945 gc_threads = NULL;
946 }
947 }
948
949 /* ----------------------------------------------------------------------------
950 Start GC threads
951 ------------------------------------------------------------------------- */
952
953 static volatile StgWord gc_running_threads;
954
955 static StgWord
956 inc_running (void)
957 {
958 StgWord new;
959 new = atomic_inc(&gc_running_threads);
960 ASSERT(new <= n_gc_threads);
961 return new;
962 }
963
964 static StgWord
965 dec_running (void)
966 {
967 ASSERT(gc_running_threads != 0);
968 return atomic_dec(&gc_running_threads);
969 }
970
971 static rtsBool
972 any_work (void)
973 {
974 int s;
975 step_workspace *ws;
976
977 gct->any_work++;
978
979 write_barrier();
980
981 // scavenge objects in compacted generation
982 if (mark_stack_bd != NULL && !mark_stack_empty()) {
983 return rtsTrue;
984 }
985
986 // Check for global work in any step. We don't need to check for
987 // local work, because we have already exited scavenge_loop(),
988 // which means there is no local work for this thread.
989 for (s = total_steps-1; s >= 0; s--) {
990 if (s == 0 && RtsFlags.GcFlags.generations > 1) {
991 continue;
992 }
993 ws = &gct->steps[s];
994 if (ws->todo_large_objects) return rtsTrue;
995 if (!looksEmptyWSDeque(ws->todo_q)) return rtsTrue;
996 if (ws->todo_overflow) return rtsTrue;
997 }
998
999 #if defined(THREADED_RTS)
1000 if (work_stealing) {
1001 nat n;
1002 // look for work to steal
1003 for (n = 0; n < n_gc_threads; n++) {
1004 if (n == gct->thread_index) continue;
1005 for (s = total_steps-1; s >= 0; s--) {
1006 ws = &gc_threads[n]->steps[s];
1007 if (!looksEmptyWSDeque(ws->todo_q)) return rtsTrue;
1008 }
1009 }
1010 }
1011 #endif
1012
1013 gct->no_work++;
1014
1015 return rtsFalse;
1016 }
1017
1018 static void
1019 scavenge_until_all_done (void)
1020 {
1021 nat r;
1022
1023
1024 loop:
1025 traceEvent(&capabilities[gct->thread_index], EVENT_GC_WORK);
1026
1027 #if defined(THREADED_RTS)
1028 if (n_gc_threads > 1) {
1029 scavenge_loop();
1030 } else {
1031 scavenge_loop1();
1032 }
1033 #else
1034 scavenge_loop();
1035 #endif
1036
1037 // scavenge_loop() only exits when there's no work to do
1038 r = dec_running();
1039
1040 traceEvent(&capabilities[gct->thread_index], EVENT_GC_IDLE);
1041
1042 debugTrace(DEBUG_gc, "%d GC threads still running", r);
1043
1044 while (gc_running_threads != 0) {
1045 // usleep(1);
1046 if (any_work()) {
1047 inc_running();
1048 goto loop;
1049 }
1050 // any_work() does not remove the work from the queue, it
1051 // just checks for the presence of work. If we find any,
1052 // then we increment gc_running_threads and go back to
1053 // scavenge_loop() to perform any pending work.
1054 }
1055
1056 traceEvent(&capabilities[gct->thread_index], EVENT_GC_DONE);
1057 }
1058
1059 #if defined(THREADED_RTS)
1060
1061 void
1062 gcWorkerThread (Capability *cap)
1063 {
1064 gc_thread *saved_gct;
1065
1066 // necessary if we stole a callee-saves register for gct:
1067 saved_gct = gct;
1068
1069 cap->in_gc = rtsTrue;
1070
1071 gct = gc_threads[cap->no];
1072 gct->id = osThreadId();
1073
1074 // Wait until we're told to wake up
1075 RELEASE_SPIN_LOCK(&gct->mut_spin);
1076 gct->wakeup = GC_THREAD_STANDING_BY;
1077 debugTrace(DEBUG_gc, "GC thread %d standing by...", gct->thread_index);
1078 ACQUIRE_SPIN_LOCK(&gct->gc_spin);
1079
1080 #ifdef USE_PAPI
1081 // start performance counters in this thread...
1082 if (gct->papi_events == -1) {
1083 papi_init_eventset(&gct->papi_events);
1084 }
1085 papi_thread_start_gc1_count(gct->papi_events);
1086 #endif
1087
1088 // Every thread evacuates some roots.
1089 gct->evac_step = 0;
1090 markSomeCapabilities(mark_root, gct, gct->thread_index, n_gc_threads,
1091 rtsTrue/*prune sparks*/);
1092 scavenge_capability_mut_lists(&capabilities[gct->thread_index]);
1093
1094 scavenge_until_all_done();
1095
1096 #ifdef USE_PAPI
1097 // count events in this thread towards the GC totals
1098 papi_thread_stop_gc1_count(gct->papi_events);
1099 #endif
1100
1101 // Wait until we're told to continue
1102 RELEASE_SPIN_LOCK(&gct->gc_spin);
1103 gct->wakeup = GC_THREAD_WAITING_TO_CONTINUE;
1104 debugTrace(DEBUG_gc, "GC thread %d waiting to continue...",
1105 gct->thread_index);
1106 ACQUIRE_SPIN_LOCK(&gct->mut_spin);
1107 debugTrace(DEBUG_gc, "GC thread %d on my way...", gct->thread_index);
1108
1109 SET_GCT(saved_gct);
1110 }
1111
1112 #endif
1113
1114 #if defined(THREADED_RTS)
1115
1116 void
1117 waitForGcThreads (Capability *cap USED_IF_THREADS)
1118 {
1119 nat n_threads = RtsFlags.ParFlags.nNodes;
1120 nat me = cap->no;
1121 nat i, j;
1122 rtsBool retry = rtsTrue;
1123
1124 while(retry) {
1125 for (i=0; i < n_threads; i++) {
1126 if (i == me) continue;
1127 if (gc_threads[i]->wakeup != GC_THREAD_STANDING_BY) {
1128 prodCapability(&capabilities[i], cap->running_task);
1129 }
1130 }
1131 for (j=0; j < 10; j++) {
1132 retry = rtsFalse;
1133 for (i=0; i < n_threads; i++) {
1134 if (i == me) continue;
1135 write_barrier();
1136 setContextSwitches();
1137 if (gc_threads[i]->wakeup != GC_THREAD_STANDING_BY) {
1138 retry = rtsTrue;
1139 }
1140 }
1141 if (!retry) break;
1142 yieldThread();
1143 }
1144 }
1145 }
1146
1147 #endif // THREADED_RTS
1148
1149 static void
1150 start_gc_threads (void)
1151 {
1152 #if defined(THREADED_RTS)
1153 gc_running_threads = 0;
1154 #endif
1155 }
1156
1157 static void
1158 wakeup_gc_threads (nat n_threads USED_IF_THREADS, nat me USED_IF_THREADS)
1159 {
1160 #if defined(THREADED_RTS)
1161 nat i;
1162 for (i=0; i < n_threads; i++) {
1163 if (i == me) continue;
1164 inc_running();
1165 debugTrace(DEBUG_gc, "waking up gc thread %d", i);
1166 if (gc_threads[i]->wakeup != GC_THREAD_STANDING_BY) barf("wakeup_gc_threads");
1167
1168 gc_threads[i]->wakeup = GC_THREAD_RUNNING;
1169 ACQUIRE_SPIN_LOCK(&gc_threads[i]->mut_spin);
1170 RELEASE_SPIN_LOCK(&gc_threads[i]->gc_spin);
1171 }
1172 #endif
1173 }
1174
1175 // After GC is complete, we must wait for all GC threads to enter the
1176 // standby state, otherwise they may still be executing inside
1177 // any_work(), and may even remain awake until the next GC starts.
1178 static void
1179 shutdown_gc_threads (nat n_threads USED_IF_THREADS, nat me USED_IF_THREADS)
1180 {
1181 #if defined(THREADED_RTS)
1182 nat i;
1183 for (i=0; i < n_threads; i++) {
1184 if (i == me) continue;
1185 while (gc_threads[i]->wakeup != GC_THREAD_WAITING_TO_CONTINUE) { write_barrier(); }
1186 }
1187 #endif
1188 }
1189
1190 #if defined(THREADED_RTS)
1191 void
1192 releaseGCThreads (Capability *cap USED_IF_THREADS)
1193 {
1194 nat n_threads = RtsFlags.ParFlags.nNodes;
1195 nat me = cap->no;
1196 nat i;
1197 for (i=0; i < n_threads; i++) {
1198 if (i == me) continue;
1199 if (gc_threads[i]->wakeup != GC_THREAD_WAITING_TO_CONTINUE)
1200 barf("releaseGCThreads");
1201
1202 gc_threads[i]->wakeup = GC_THREAD_INACTIVE;
1203 ACQUIRE_SPIN_LOCK(&gc_threads[i]->gc_spin);
1204 RELEASE_SPIN_LOCK(&gc_threads[i]->mut_spin);
1205 }
1206 }
1207 #endif
1208
1209 /* ----------------------------------------------------------------------------
1210 Initialise a generation that is to be collected
1211 ------------------------------------------------------------------------- */
1212
1213 static void
1214 init_collected_gen (nat g, nat n_threads)
1215 {
1216 nat s, t, i;
1217 step_workspace *ws;
1218 step *stp;
1219 bdescr *bd;
1220
1221 // Throw away the current mutable list. Invariant: the mutable
1222 // list always has at least one block; this means we can avoid a
1223 // check for NULL in recordMutable().
1224 if (g != 0) {
1225 freeChain(generations[g].mut_list);
1226 generations[g].mut_list = allocBlock();
1227 for (i = 0; i < n_capabilities; i++) {
1228 freeChain(capabilities[i].mut_lists[g]);
1229 capabilities[i].mut_lists[g] = allocBlock();
1230 }
1231 }
1232
1233 for (s = 0; s < generations[g].n_steps; s++) {
1234
1235 stp = &generations[g].steps[s];
1236 ASSERT(stp->gen_no == g);
1237
1238 // we'll construct a new list of threads in this step
1239 // during GC, throw away the current list.
1240 stp->old_threads = stp->threads;
1241 stp->threads = END_TSO_QUEUE;
1242
1243 // generation 0, step 0 doesn't need to-space
1244 if (g == 0 && s == 0 && RtsFlags.GcFlags.generations > 1) {
1245 continue;
1246 }
1247
1248 // deprecate the existing blocks
1249 stp->old_blocks = stp->blocks;
1250 stp->n_old_blocks = stp->n_blocks;
1251 stp->blocks = NULL;
1252 stp->n_blocks = 0;
1253 stp->n_words = 0;
1254 stp->live_estimate = 0;
1255
1256 // initialise the large object queues.
1257 stp->scavenged_large_objects = NULL;
1258 stp->n_scavenged_large_blocks = 0;
1259
1260 // mark the small objects as from-space
1261 for (bd = stp->old_blocks; bd; bd = bd->link) {
1262 bd->flags &= ~BF_EVACUATED;
1263 }
1264
1265 // mark the large objects as from-space
1266 for (bd = stp->large_objects; bd; bd = bd->link) {
1267 bd->flags &= ~BF_EVACUATED;
1268 }
1269
1270 // for a compacted step, we need to allocate the bitmap
1271 if (stp->mark) {
1272 nat bitmap_size; // in bytes
1273 bdescr *bitmap_bdescr;
1274 StgWord *bitmap;
1275
1276 bitmap_size = stp->n_old_blocks * BLOCK_SIZE / (sizeof(W_)*BITS_PER_BYTE);
1277
1278 if (bitmap_size > 0) {
1279 bitmap_bdescr = allocGroup((lnat)BLOCK_ROUND_UP(bitmap_size)
1280 / BLOCK_SIZE);
1281 stp->bitmap = bitmap_bdescr;
1282 bitmap = bitmap_bdescr->start;
1283
1284 debugTrace(DEBUG_gc, "bitmap_size: %d, bitmap: %p",
1285 bitmap_size, bitmap);
1286
1287 // don't forget to fill it with zeros!
1288 memset(bitmap, 0, bitmap_size);
1289
1290 // For each block in this step, point to its bitmap from the
1291 // block descriptor.
1292 for (bd=stp->old_blocks; bd != NULL; bd = bd->link) {
1293 bd->u.bitmap = bitmap;
1294 bitmap += BLOCK_SIZE_W / (sizeof(W_)*BITS_PER_BYTE);
1295
1296 // Also at this point we set the BF_MARKED flag
1297 // for this block. The invariant is that
1298 // BF_MARKED is always unset, except during GC
1299 // when it is set on those blocks which will be
1300 // compacted.
1301 if (!(bd->flags & BF_FRAGMENTED)) {
1302 bd->flags |= BF_MARKED;
1303 }
1304 }
1305 }
1306 }
1307 }
1308
1309 // For each GC thread, for each step, allocate a "todo" block to
1310 // store evacuated objects to be scavenged, and a block to store
1311 // evacuated objects that do not need to be scavenged.
1312 for (t = 0; t < n_threads; t++) {
1313 for (s = 0; s < generations[g].n_steps; s++) {
1314
1315 // we don't copy objects into g0s0, unless -G0
1316 if (g==0 && s==0 && RtsFlags.GcFlags.generations > 1) continue;
1317
1318 ws = &gc_threads[t]->steps[g * RtsFlags.GcFlags.steps + s];
1319
1320 ws->todo_large_objects = NULL;
1321
1322 ws->part_list = NULL;
1323 ws->n_part_blocks = 0;
1324
1325 // allocate the first to-space block; extra blocks will be
1326 // chained on as necessary.
1327 ws->todo_bd = NULL;
1328 ASSERT(looksEmptyWSDeque(ws->todo_q));
1329 alloc_todo_block(ws,0);
1330
1331 ws->todo_overflow = NULL;
1332 ws->n_todo_overflow = 0;
1333
1334 ws->scavd_list = NULL;
1335 ws->n_scavd_blocks = 0;
1336 }
1337 }
1338 }
1339
1340
1341 /* ----------------------------------------------------------------------------
1342 Initialise a generation that is *not* to be collected
1343 ------------------------------------------------------------------------- */
1344
1345 static void
1346 init_uncollected_gen (nat g, nat threads)
1347 {
1348 nat s, t, n;
1349 step_workspace *ws;
1350 step *stp;
1351 bdescr *bd;
1352
1353 // save the current mutable lists for this generation, and
1354 // allocate a fresh block for each one. We'll traverse these
1355 // mutable lists as roots early on in the GC.
1356 generations[g].saved_mut_list = generations[g].mut_list;
1357 generations[g].mut_list = allocBlock();
1358 for (n = 0; n < n_capabilities; n++) {
1359 capabilities[n].saved_mut_lists[g] = capabilities[n].mut_lists[g];
1360 capabilities[n].mut_lists[g] = allocBlock();
1361 }
1362
1363 for (s = 0; s < generations[g].n_steps; s++) {
1364 stp = &generations[g].steps[s];
1365 stp->scavenged_large_objects = NULL;
1366 stp->n_scavenged_large_blocks = 0;
1367 }
1368
1369 for (s = 0; s < generations[g].n_steps; s++) {
1370
1371 stp = &generations[g].steps[s];
1372
1373 for (t = 0; t < threads; t++) {
1374 ws = &gc_threads[t]->steps[g * RtsFlags.GcFlags.steps + s];
1375
1376 ASSERT(looksEmptyWSDeque(ws->todo_q));
1377 ws->todo_large_objects = NULL;
1378
1379 ws->part_list = NULL;
1380 ws->n_part_blocks = 0;
1381
1382 ws->scavd_list = NULL;
1383 ws->n_scavd_blocks = 0;
1384
1385 // If the block at the head of the list in this generation
1386 // is less than 3/4 full, then use it as a todo block.
1387 if (stp->blocks && isPartiallyFull(stp->blocks))
1388 {
1389 ws->todo_bd = stp->blocks;
1390 ws->todo_free = ws->todo_bd->free;
1391 ws->todo_lim = ws->todo_bd->start + BLOCK_SIZE_W;
1392 stp->blocks = stp->blocks->link;
1393 stp->n_blocks -= 1;
1394 stp->n_words -= ws->todo_bd->free - ws->todo_bd->start;
1395 ws->todo_bd->link = NULL;
1396 // we must scan from the current end point.
1397 ws->todo_bd->u.scan = ws->todo_bd->free;
1398 }
1399 else
1400 {
1401 ws->todo_bd = NULL;
1402 alloc_todo_block(ws,0);
1403 }
1404 }
1405
1406 // deal out any more partial blocks to the threads' part_lists
1407 t = 0;
1408 while (stp->blocks && isPartiallyFull(stp->blocks))
1409 {
1410 bd = stp->blocks;
1411 stp->blocks = bd->link;
1412 ws = &gc_threads[t]->steps[g * RtsFlags.GcFlags.steps + s];
1413 bd->link = ws->part_list;
1414 ws->part_list = bd;
1415 ws->n_part_blocks += 1;
1416 bd->u.scan = bd->free;
1417 stp->n_blocks -= 1;
1418 stp->n_words -= bd->free - bd->start;
1419 t++;
1420 if (t == n_gc_threads) t = 0;
1421 }
1422 }
1423 }
1424
1425 /* -----------------------------------------------------------------------------
1426 Initialise a gc_thread before GC
1427 -------------------------------------------------------------------------- */
1428
1429 static void
1430 init_gc_thread (gc_thread *t)
1431 {
1432 t->static_objects = END_OF_STATIC_LIST;
1433 t->scavenged_static_objects = END_OF_STATIC_LIST;
1434 t->scan_bd = NULL;
1435 t->mut_lists = capabilities[t->thread_index].mut_lists;
1436 t->evac_step = 0;
1437 t->failed_to_evac = rtsFalse;
1438 t->eager_promotion = rtsTrue;
1439 t->thunk_selector_depth = 0;
1440 t->copied = 0;
1441 t->scanned = 0;
1442 t->any_work = 0;
1443 t->no_work = 0;
1444 t->scav_find_work = 0;
1445 }
1446
1447 /* -----------------------------------------------------------------------------
1448 Function we pass to evacuate roots.
1449 -------------------------------------------------------------------------- */
1450
1451 static void
1452 mark_root(void *user USED_IF_THREADS, StgClosure **root)
1453 {
1454 // we stole a register for gct, but this function is called from
1455 // *outside* the GC where the register variable is not in effect,
1456 // so we need to save and restore it here. NB. only call
1457 // mark_root() from the main GC thread, otherwise gct will be
1458 // incorrect.
1459 gc_thread *saved_gct;
1460 saved_gct = gct;
1461 SET_GCT(user);
1462
1463 evacuate(root);
1464
1465 SET_GCT(saved_gct);
1466 }
1467
1468 /* -----------------------------------------------------------------------------
1469 Initialising the static object & mutable lists
1470 -------------------------------------------------------------------------- */
1471
1472 static void
1473 zero_static_object_list(StgClosure* first_static)
1474 {
1475 StgClosure* p;
1476 StgClosure* link;
1477 const StgInfoTable *info;
1478
1479 for (p = first_static; p != END_OF_STATIC_LIST; p = link) {
1480 info = get_itbl(p);
1481 link = *STATIC_LINK(info, p);
1482 *STATIC_LINK(info,p) = NULL;
1483 }
1484 }
1485
1486 /* ----------------------------------------------------------------------------
1487 Update the pointers from the task list
1488
1489 These are treated as weak pointers because we want to allow a main
1490 thread to get a BlockedOnDeadMVar exception in the same way as any
1491 other thread. Note that the threads should all have been retained
1492 by GC by virtue of being on the all_threads list, we're just
1493 updating pointers here.
1494 ------------------------------------------------------------------------- */
1495
1496 static void
1497 update_task_list (void)
1498 {
1499 Task *task;
1500 StgTSO *tso;
1501 for (task = all_tasks; task != NULL; task = task->all_link) {
1502 if (!task->stopped && task->tso) {
1503 ASSERT(task->tso->bound == task);
1504 tso = (StgTSO *) isAlive((StgClosure *)task->tso);
1505 if (tso == NULL) {
1506 barf("task %p: main thread %d has been GC'd",
1507 #ifdef THREADED_RTS
1508 (void *)task->id,
1509 #else
1510 (void *)task,
1511 #endif
1512 task->tso->id);
1513 }
1514 task->tso = tso;
1515 }
1516 }
1517 }
1518
1519 /* ----------------------------------------------------------------------------
1520 Reset the sizes of the older generations when we do a major
1521 collection.
1522
1523 CURRENT STRATEGY: make all generations except zero the same size.
1524 We have to stay within the maximum heap size, and leave a certain
1525 percentage of the maximum heap size available to allocate into.
1526 ------------------------------------------------------------------------- */
1527
1528 static void
1529 resize_generations (void)
1530 {
1531 nat g;
1532
1533 if (major_gc && RtsFlags.GcFlags.generations > 1) {
1534 nat live, size, min_alloc, words;
1535 nat max = RtsFlags.GcFlags.maxHeapSize;
1536 nat gens = RtsFlags.GcFlags.generations;
1537
1538 // live in the oldest generations
1539 if (oldest_gen->steps[0].live_estimate != 0) {
1540 words = oldest_gen->steps[0].live_estimate;
1541 } else {
1542 words = oldest_gen->steps[0].n_words;
1543 }
1544 live = (words + BLOCK_SIZE_W - 1) / BLOCK_SIZE_W +
1545 oldest_gen->steps[0].n_large_blocks;
1546
1547 // default max size for all generations except zero
1548 size = stg_max(live * RtsFlags.GcFlags.oldGenFactor,
1549 RtsFlags.GcFlags.minOldGenSize);
1550
1551 // minimum size for generation zero
1552 min_alloc = stg_max((RtsFlags.GcFlags.pcFreeHeap * max) / 200,
1553 RtsFlags.GcFlags.minAllocAreaSize);
1554
1555 // Auto-enable compaction when the residency reaches a
1556 // certain percentage of the maximum heap size (default: 30%).
1557 if (RtsFlags.GcFlags.generations > 1 &&
1558 (RtsFlags.GcFlags.compact ||
1559 (max > 0 &&
1560 oldest_gen->steps[0].n_blocks >
1561 (RtsFlags.GcFlags.compactThreshold * max) / 100))) {
1562 oldest_gen->steps[0].mark = 1;
1563 oldest_gen->steps[0].compact = 1;
1564 // debugBelch("compaction: on\n", live);
1565 } else {
1566 oldest_gen->steps[0].mark = 0;
1567 oldest_gen->steps[0].compact = 0;
1568 // debugBelch("compaction: off\n", live);
1569 }
1570
1571 if (RtsFlags.GcFlags.sweep) {
1572 oldest_gen->steps[0].mark = 1;
1573 }
1574
1575 // if we're going to go over the maximum heap size, reduce the
1576 // size of the generations accordingly. The calculation is
1577 // different if compaction is turned on, because we don't need
1578 // to double the space required to collect the old generation.
1579 if (max != 0) {
1580
1581 // this test is necessary to ensure that the calculations
1582 // below don't have any negative results - we're working
1583 // with unsigned values here.
1584 if (max < min_alloc) {
1585 heapOverflow();
1586 }
1587
1588 if (oldest_gen->steps[0].compact) {
1589 if ( (size + (size - 1) * (gens - 2) * 2) + min_alloc > max ) {
1590 size = (max - min_alloc) / ((gens - 1) * 2 - 1);
1591 }
1592 } else {
1593 if ( (size * (gens - 1) * 2) + min_alloc > max ) {
1594 size = (max - min_alloc) / ((gens - 1) * 2);
1595 }
1596 }
1597
1598 if (size < live) {
1599 heapOverflow();
1600 }
1601 }
1602
1603 #if 0
1604 debugBelch("live: %d, min_alloc: %d, size : %d, max = %d\n", live,
1605 min_alloc, size, max);
1606 #endif
1607
1608 for (g = 0; g < gens; g++) {
1609 generations[g].max_blocks = size;
1610 }
1611 }
1612 }
1613
1614 /* -----------------------------------------------------------------------------
1615 Calculate the new size of the nursery, and resize it.
1616 -------------------------------------------------------------------------- */
1617
1618 static void
1619 resize_nursery (void)
1620 {
1621 lnat min_nursery = RtsFlags.GcFlags.minAllocAreaSize * n_capabilities;
1622
1623 if (RtsFlags.GcFlags.generations == 1)
1624 { // Two-space collector:
1625 nat blocks;
1626
1627 /* set up a new nursery. Allocate a nursery size based on a
1628 * function of the amount of live data (by default a factor of 2)
1629 * Use the blocks from the old nursery if possible, freeing up any
1630 * left over blocks.
1631 *
1632 * If we get near the maximum heap size, then adjust our nursery
1633 * size accordingly. If the nursery is the same size as the live
1634 * data (L), then we need 3L bytes. We can reduce the size of the
1635 * nursery to bring the required memory down near 2L bytes.
1636 *
1637 * A normal 2-space collector would need 4L bytes to give the same
1638 * performance we get from 3L bytes, reducing to the same
1639 * performance at 2L bytes.
1640 */
1641 blocks = g0s0->n_blocks;
1642
1643 if ( RtsFlags.GcFlags.maxHeapSize != 0 &&
1644 blocks * RtsFlags.GcFlags.oldGenFactor * 2 >
1645 RtsFlags.GcFlags.maxHeapSize )
1646 {
1647 long adjusted_blocks; // signed on purpose
1648 int pc_free;
1649
1650 adjusted_blocks = (RtsFlags.GcFlags.maxHeapSize - 2 * blocks);
1651
1652 debugTrace(DEBUG_gc, "near maximum heap size of 0x%x blocks, blocks = %d, adjusted to %ld",
1653 RtsFlags.GcFlags.maxHeapSize, blocks, adjusted_blocks);
1654
1655 pc_free = adjusted_blocks * 100 / RtsFlags.GcFlags.maxHeapSize;
1656 if (pc_free < RtsFlags.GcFlags.pcFreeHeap) /* might even * be < 0 */
1657 {
1658 heapOverflow();
1659 }
1660 blocks = adjusted_blocks;
1661 }
1662 else
1663 {
1664 blocks *= RtsFlags.GcFlags.oldGenFactor;
1665 if (blocks < min_nursery)
1666 {
1667 blocks = min_nursery;
1668 }
1669 }
1670 resizeNurseries(blocks);
1671 }
1672 else // Generational collector
1673 {
1674 /*
1675 * If the user has given us a suggested heap size, adjust our
1676 * allocation area to make best use of the memory available.
1677 */
1678 if (RtsFlags.GcFlags.heapSizeSuggestion)
1679 {
1680 long blocks;
1681 nat needed = calcNeeded(); // approx blocks needed at next GC
1682
1683 /* Guess how much will be live in generation 0 step 0 next time.
1684 * A good approximation is obtained by finding the
1685 * percentage of g0s0 that was live at the last minor GC.
1686 *
1687 * We have an accurate figure for the amount of copied data in
1688 * 'copied', but we must convert this to a number of blocks, with
1689 * a small adjustment for estimated slop at the end of a block
1690 * (- 10 words).
1691 */
1692 if (N == 0)
1693 {
1694 g0s0_pcnt_kept = ((copied / (BLOCK_SIZE_W - 10)) * 100)
1695 / countNurseryBlocks();
1696 }
1697
1698 /* Estimate a size for the allocation area based on the
1699 * information available. We might end up going slightly under
1700 * or over the suggested heap size, but we should be pretty
1701 * close on average.
1702 *
1703 * Formula: suggested - needed
1704 * ----------------------------
1705 * 1 + g0s0_pcnt_kept/100
1706 *
1707 * where 'needed' is the amount of memory needed at the next
1708 * collection for collecting all steps except g0s0.
1709 */
1710 blocks =
1711 (((long)RtsFlags.GcFlags.heapSizeSuggestion - (long)needed) * 100) /
1712 (100 + (long)g0s0_pcnt_kept);
1713
1714 if (blocks < (long)min_nursery) {
1715 blocks = min_nursery;
1716 }
1717
1718 resizeNurseries((nat)blocks);
1719 }
1720 else
1721 {
1722 // we might have added extra large blocks to the nursery, so
1723 // resize back to minAllocAreaSize again.
1724 resizeNurseriesFixed(RtsFlags.GcFlags.minAllocAreaSize);
1725 }
1726 }
1727 }
1728
1729 /* -----------------------------------------------------------------------------
1730 Sanity code for CAF garbage collection.
1731
1732 With DEBUG turned on, we manage a CAF list in addition to the SRT
1733 mechanism. After GC, we run down the CAF list and blackhole any
1734 CAFs which have been garbage collected. This means we get an error
1735 whenever the program tries to enter a garbage collected CAF.
1736
1737 Any garbage collected CAFs are taken off the CAF list at the same
1738 time.
1739 -------------------------------------------------------------------------- */
1740
1741 #if 0 && defined(DEBUG)
1742
1743 static void
1744 gcCAFs(void)
1745 {
1746 StgClosure* p;
1747 StgClosure** pp;
1748 const StgInfoTable *info;
1749 nat i;
1750
1751 i = 0;
1752 p = caf_list;
1753 pp = &caf_list;
1754
1755 while (p != NULL) {
1756
1757 info = get_itbl(p);
1758
1759 ASSERT(info->type == IND_STATIC);
1760
1761 if (STATIC_LINK(info,p) == NULL) {
1762 debugTrace(DEBUG_gccafs, "CAF gc'd at 0x%04lx", (long)p);
1763 // black hole it
1764 SET_INFO(p,&stg_BLACKHOLE_info);
1765 p = STATIC_LINK2(info,p);
1766 *pp = p;
1767 }
1768 else {
1769 pp = &STATIC_LINK2(info,p);
1770 p = *pp;
1771 i++;
1772 }
1773
1774 }
1775
1776 debugTrace(DEBUG_gccafs, "%d CAFs live", i);
1777 }
1778 #endif