1 /* -----------------------------------------------------------------------------
2 * $Id: Storage.c,v 1.33 2001/01/24 15:46:19 simonmar Exp $
4 * (c) The GHC Team, 1998-1999
6 * Storage manager front end
8 * ---------------------------------------------------------------------------*/
15 #include "BlockAlloc.h"
22 #include "StoragePriv.h"
25 nat nursery_blocks; /* number of blocks in the nursery */
28 StgClosure *caf_list = NULL;
30 bdescr *small_alloc_list; /* allocate()d small objects */
31 bdescr *large_alloc_list; /* allocate()d large objects */
32 nat alloc_blocks; /* number of allocate()d blocks since GC */
33 nat alloc_blocks_lim; /* approximate limit on alloc_blocks */
35 StgPtr alloc_Hp = NULL; /* next free byte in small_alloc_list */
36 StgPtr alloc_HpLim = NULL; /* end of block at small_alloc_list */
38 generation *generations; /* all the generations */
39 generation *g0; /* generation 0, for convenience */
40 generation *oldest_gen; /* oldest generation, for convenience */
41 step *g0s0; /* generation 0, step 0, for convenience */
43 lnat total_allocated = 0; /* total memory allocated during run */
46 * Storage manager mutex: protects all the above state from
47 * simultaneous access by two STG threads.
50 pthread_mutex_t sm_mutex = PTHREAD_MUTEX_INITIALIZER;
56 static void *stgAllocForGMP (size_t size_in_bytes);
57 static void *stgReallocForGMP (void *ptr, size_t old_size, size_t new_size);
58 static void stgDeallocForGMP (void *ptr, size_t size);
67 /* If we're doing heap profiling, we want a two-space heap with a
68 * fixed-size allocation area so that we get roughly even-spaced
71 #if defined(PROFILING) || defined(DEBUG)
72 if (RtsFlags.ProfFlags.doHeapProfile) {
73 RtsFlags.GcFlags.generations = 1;
74 RtsFlags.GcFlags.steps = 1;
75 RtsFlags.GcFlags.oldGenFactor = 0;
76 RtsFlags.GcFlags.heapSizeSuggestion = 0;
80 if (RtsFlags.GcFlags.heapSizeSuggestion >
81 RtsFlags.GcFlags.maxHeapSize) {
82 RtsFlags.GcFlags.maxHeapSize = RtsFlags.GcFlags.heapSizeSuggestion;
87 /* allocate generation info array */
88 generations = (generation *)stgMallocBytes(RtsFlags.GcFlags.generations
89 * sizeof(struct _generation),
92 /* Initialise all generations */
93 for(g = 0; g < RtsFlags.GcFlags.generations; g++) {
94 gen = &generations[g];
96 gen->mut_list = END_MUT_LIST;
97 gen->mut_once_list = END_MUT_LIST;
99 gen->failed_promotions = 0;
103 /* A couple of convenience pointers */
104 g0 = &generations[0];
105 oldest_gen = &generations[RtsFlags.GcFlags.generations-1];
107 /* Allocate step structures in each generation */
108 if (RtsFlags.GcFlags.generations > 1) {
109 /* Only for multiple-generations */
111 /* Oldest generation: one step */
112 oldest_gen->n_steps = 1;
114 stgMallocBytes(1 * sizeof(struct _step), "initStorage: last step");
116 /* set up all except the oldest generation with 2 steps */
117 for(g = 0; g < RtsFlags.GcFlags.generations-1; g++) {
118 generations[g].n_steps = RtsFlags.GcFlags.steps;
119 generations[g].steps =
120 stgMallocBytes (RtsFlags.GcFlags.steps * sizeof(struct _step),
121 "initStorage: steps");
125 /* single generation, i.e. a two-space collector */
127 g0->steps = stgMallocBytes (sizeof(struct _step), "initStorage: steps");
130 /* Initialise all steps */
131 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
132 for (s = 0; s < generations[g].n_steps; s++) {
133 stp = &generations[g].steps[s];
137 stp->gen = &generations[g];
143 stp->large_objects = NULL;
144 stp->new_large_objects = NULL;
145 stp->scavenged_large_objects = NULL;
149 /* Set up the destination pointers in each younger gen. step */
150 for (g = 0; g < RtsFlags.GcFlags.generations-1; g++) {
151 for (s = 0; s < generations[g].n_steps-1; s++) {
152 generations[g].steps[s].to = &generations[g].steps[s+1];
154 generations[g].steps[s].to = &generations[g+1].steps[0];
157 /* The oldest generation has one step and its destination is the
159 oldest_gen->steps[0].to = &oldest_gen->steps[0];
161 /* generation 0 is special: that's the nursery */
162 generations[0].max_blocks = 0;
164 /* G0S0: the allocation area. Policy: keep the allocation area
165 * small to begin with, even if we have a large suggested heap
166 * size. Reason: we're going to do a major collection first, and we
167 * don't want it to be a big one. This vague idea is borne out by
168 * rigorous experimental evidence.
170 g0s0 = &generations[0].steps[0];
174 weak_ptr_list = NULL;
177 /* initialise the allocate() interface */
178 small_alloc_list = NULL;
179 large_alloc_list = NULL;
181 alloc_blocks_lim = RtsFlags.GcFlags.minAllocAreaSize;
183 /* Tell GNU multi-precision pkg about our custom alloc functions */
184 mp_set_memory_functions(stgAllocForGMP, stgReallocForGMP, stgDeallocForGMP);
187 pthread_mutex_init(&sm_mutex, NULL);
190 IF_DEBUG(gc, stat_describe_gens());
196 stat_exit(calcAllocated());
199 /* -----------------------------------------------------------------------------
201 -------------------------------------------------------------------------- */
204 newCAF(StgClosure* caf)
206 /* Put this CAF on the mutable list for the old generation.
207 * This is a HACK - the IND_STATIC closure doesn't really have
208 * a mut_link field, but we pretend it has - in fact we re-use
209 * the STATIC_LINK field for the time being, because when we
210 * come to do a major GC we won't need the mut_link field
211 * any more and can use it as a STATIC_LINK.
213 ACQUIRE_LOCK(&sm_mutex);
215 ASSERT( ((StgMutClosure*)caf)->mut_link == NULL );
216 ((StgMutClosure *)caf)->mut_link = oldest_gen->mut_once_list;
217 oldest_gen->mut_once_list = (StgMutClosure *)caf;
220 /* For dynamically-loaded code, we retain all the CAFs. There is no
221 * way of knowing which ones we'll need in the future.
223 if (is_dynamically_loaded_rwdata_ptr((StgPtr)caf)) {
224 caf->payload[2] = caf_list; /* IND_STATIC_LINK2() */
230 /* If we're Hugs, we also have to put it in the CAF table, so that
231 the CAF can be reverted. When reverting, CAFs created by compiled
232 code are recorded in the CAF table, which lives outside the
233 heap, in mallocville. CAFs created by interpreted code are
234 chained together via the link fields in StgCAFs, and are not
235 recorded in the CAF table.
237 ASSERT( get_itbl(caf)->type == THUNK_STATIC );
238 addToECafTable ( caf, get_itbl(caf) );
241 RELEASE_LOCK(&sm_mutex);
250 for (p = caf_list; p != NULL; p = STATIC_LINK2(get_itbl(p),p)) {
258 newCAF_made_by_Hugs(StgCAF* caf)
260 ACQUIRE_LOCK(&sm_mutex);
262 ASSERT( get_itbl(caf)->type == CAF_ENTERED );
263 recordOldToNewPtrs((StgMutClosure*)caf);
264 caf->link = ecafList;
265 ecafList = caf->link;
267 RELEASE_LOCK(&sm_mutex);
272 /* These initialisations are critical for correct operation
273 on the first call of addToECafTable.
275 StgCAF* ecafList = END_ECAF_LIST;
276 StgCAFTabEntry* ecafTable = NULL;
277 StgInt usedECafTable = 0;
278 StgInt sizeECafTable = 0;
281 void clearECafTable ( void )
286 void addToECafTable ( StgClosure* closure, StgInfoTable* origItbl )
290 if (usedECafTable == sizeECafTable) {
291 /* Make the initial table size be 8 */
293 if (sizeECafTable == 0) sizeECafTable = 8;
294 et2 = stgMallocBytes (
295 sizeECafTable * sizeof(StgCAFTabEntry),
297 for (i = 0; i < usedECafTable; i++)
298 et2[i] = ecafTable[i];
299 if (ecafTable) free(ecafTable);
302 ecafTable[usedECafTable].closure = closure;
303 ecafTable[usedECafTable].origItbl = origItbl;
308 /* -----------------------------------------------------------------------------
310 -------------------------------------------------------------------------- */
313 allocNurseries( void )
322 for (cap = free_capabilities; cap != NULL; cap = cap->link) {
323 cap->rNursery = allocNursery(NULL, RtsFlags.GcFlags.minAllocAreaSize);
324 cap->rCurrentNursery = cap->rNursery;
325 for (bd = cap->rNursery; bd != NULL; bd = bd->link) {
326 bd->back = (bdescr *)cap;
329 /* Set the back links to be equal to the Capability,
330 * so we can do slightly better informed locking.
334 nursery_blocks = RtsFlags.GcFlags.minAllocAreaSize;
335 g0s0->blocks = allocNursery(NULL, nursery_blocks);
336 g0s0->n_blocks = nursery_blocks;
337 g0s0->to_space = NULL;
338 MainRegTable.rNursery = g0s0->blocks;
339 MainRegTable.rCurrentNursery = g0s0->blocks;
340 /* hp, hpLim, hp_bd, to_space etc. aren't used in G0S0 */
345 resetNurseries( void )
351 /* All tasks must be stopped */
352 ASSERT(n_free_capabilities == RtsFlags.ParFlags.nNodes);
354 for (cap = free_capabilities; cap != NULL; cap = cap->link) {
355 for (bd = cap->rNursery; bd; bd = bd->link) {
356 bd->free = bd->start;
357 ASSERT(bd->gen == g0);
358 ASSERT(bd->step == g0s0);
359 IF_DEBUG(sanity,memset(bd->start, 0xaa, BLOCK_SIZE));
361 cap->rCurrentNursery = cap->rNursery;
364 for (bd = g0s0->blocks; bd; bd = bd->link) {
365 bd->free = bd->start;
366 ASSERT(bd->gen == g0);
367 ASSERT(bd->step == g0s0);
368 IF_DEBUG(sanity,memset(bd->start, 0xaa, BLOCK_SIZE));
370 MainRegTable.rNursery = g0s0->blocks;
371 MainRegTable.rCurrentNursery = g0s0->blocks;
376 allocNursery (bdescr *last_bd, nat blocks)
381 /* Allocate a nursery */
382 for (i=0; i < blocks; i++) {
388 bd->free = bd->start;
395 resizeNursery ( nat blocks )
400 barf("resizeNursery: can't resize in SMP mode");
403 if (nursery_blocks == blocks) {
404 ASSERT(g0s0->n_blocks == blocks);
408 else if (nursery_blocks < blocks) {
409 IF_DEBUG(gc, fprintf(stderr, "Increasing size of nursery to %d blocks\n",
411 g0s0->blocks = allocNursery(g0s0->blocks, blocks-nursery_blocks);
417 IF_DEBUG(gc, fprintf(stderr, "Decreasing size of nursery to %d blocks\n",
419 for (bd = g0s0->blocks; nursery_blocks > blocks; nursery_blocks--) {
427 g0s0->n_blocks = nursery_blocks = blocks;
430 /* -----------------------------------------------------------------------------
431 The allocate() interface
433 allocate(n) always succeeds, and returns a chunk of memory n words
434 long. n can be larger than the size of a block if necessary, in
435 which case a contiguous block group will be allocated.
436 -------------------------------------------------------------------------- */
444 ACQUIRE_LOCK(&sm_mutex);
446 TICK_ALLOC_HEAP_NOCTR(n);
449 /* big allocation (>LARGE_OBJECT_THRESHOLD) */
450 /* ToDo: allocate directly into generation 1 */
451 if (n >= LARGE_OBJECT_THRESHOLD/sizeof(W_)) {
452 nat req_blocks = (lnat)BLOCK_ROUND_UP(n*sizeof(W_)) / BLOCK_SIZE;
453 bd = allocGroup(req_blocks);
454 dbl_link_onto(bd, &g0s0->large_objects);
458 bd->free = bd->start;
459 /* don't add these blocks to alloc_blocks, since we're assuming
460 * that large objects are likely to remain live for quite a while
461 * (eg. running threads), so garbage collecting early won't make
464 alloc_blocks += req_blocks;
465 RELEASE_LOCK(&sm_mutex);
468 /* small allocation (<LARGE_OBJECT_THRESHOLD) */
469 } else if (small_alloc_list == NULL || alloc_Hp + n > alloc_HpLim) {
470 if (small_alloc_list) {
471 small_alloc_list->free = alloc_Hp;
474 bd->link = small_alloc_list;
475 small_alloc_list = bd;
479 alloc_Hp = bd->start;
480 alloc_HpLim = bd->start + BLOCK_SIZE_W;
486 RELEASE_LOCK(&sm_mutex);
490 lnat allocated_bytes(void)
492 return (alloc_blocks * BLOCK_SIZE_W - (alloc_HpLim - alloc_Hp));
495 /* -----------------------------------------------------------------------------
496 Allocation functions for GMP.
498 These all use the allocate() interface - we can't have any garbage
499 collection going on during a gmp operation, so we use allocate()
500 which always succeeds. The gmp operations which might need to
501 allocate will ask the storage manager (via doYouWantToGC()) whether
502 a garbage collection is required, in case we get into a loop doing
503 only allocate() style allocation.
504 -------------------------------------------------------------------------- */
507 stgAllocForGMP (size_t size_in_bytes)
510 nat data_size_in_words, total_size_in_words;
512 /* should be a multiple of sizeof(StgWord) (whole no. of limbs) */
513 ASSERT(size_in_bytes % sizeof(W_) == 0);
515 data_size_in_words = size_in_bytes / sizeof(W_);
516 total_size_in_words = sizeofW(StgArrWords) + data_size_in_words;
518 /* allocate and fill it in. */
519 arr = (StgArrWords *)allocate(total_size_in_words);
520 SET_ARR_HDR(arr, &stg_ARR_WORDS_info, CCCS, data_size_in_words);
522 /* and return a ptr to the goods inside the array */
523 return(BYTE_ARR_CTS(arr));
527 stgReallocForGMP (void *ptr, size_t old_size, size_t new_size)
529 void *new_stuff_ptr = stgAllocForGMP(new_size);
531 char *p = (char *) ptr;
532 char *q = (char *) new_stuff_ptr;
534 for (; i < old_size; i++, p++, q++) {
538 return(new_stuff_ptr);
542 stgDeallocForGMP (void *ptr STG_UNUSED,
543 size_t size STG_UNUSED)
545 /* easy for us: the garbage collector does the dealloc'n */
548 /* -----------------------------------------------------------------------------
550 * -------------------------------------------------------------------------- */
552 /* -----------------------------------------------------------------------------
555 * Approximate how much we've allocated: number of blocks in the
556 * nursery + blocks allocated via allocate() - unused nusery blocks.
557 * This leaves a little slop at the end of each block, and doesn't
558 * take into account large objects (ToDo).
559 * -------------------------------------------------------------------------- */
562 calcAllocated( void )
570 /* All tasks must be stopped. Can't assert that all the
571 capabilities are owned by the scheduler, though: one or more
572 tasks might have been stopped while they were running (non-main)
574 /* ASSERT(n_free_capabilities == RtsFlags.ParFlags.nNodes); */
577 n_free_capabilities * RtsFlags.GcFlags.minAllocAreaSize * BLOCK_SIZE_W
580 for (cap = free_capabilities; cap != NULL; cap = cap->link) {
581 for ( bd = cap->rCurrentNursery->link; bd != NULL; bd = bd->link ) {
582 allocated -= BLOCK_SIZE_W;
584 if (cap->rCurrentNursery->free < cap->rCurrentNursery->start
586 allocated -= (cap->rCurrentNursery->start + BLOCK_SIZE_W)
587 - cap->rCurrentNursery->free;
592 bdescr *current_nursery = MainRegTable.rCurrentNursery;
594 allocated = (nursery_blocks * BLOCK_SIZE_W) + allocated_bytes();
595 for ( bd = current_nursery->link; bd != NULL; bd = bd->link ) {
596 allocated -= BLOCK_SIZE_W;
598 if (current_nursery->free < current_nursery->start + BLOCK_SIZE_W) {
599 allocated -= (current_nursery->start + BLOCK_SIZE_W)
600 - current_nursery->free;
604 total_allocated += allocated;
608 /* Approximate the amount of live data in the heap. To be called just
609 * after garbage collection (see GarbageCollect()).
618 if (RtsFlags.GcFlags.generations == 1) {
619 live = (g0s0->to_blocks - 1) * BLOCK_SIZE_W +
620 ((lnat)g0s0->hp_bd->free - (lnat)g0s0->hp_bd->start) / sizeof(W_);
624 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
625 for (s = 0; s < generations[g].n_steps; s++) {
626 /* approximate amount of live data (doesn't take into account slop
627 * at end of each block).
629 if (g == 0 && s == 0) {
632 stp = &generations[g].steps[s];
633 live += (stp->n_blocks - 1) * BLOCK_SIZE_W +
634 ((lnat)stp->hp_bd->free - (lnat)stp->hp_bd->start) / sizeof(W_);
640 /* Approximate the number of blocks that will be needed at the next
641 * garbage collection.
643 * Assume: all data currently live will remain live. Steps that will
644 * be collected next time will therefore need twice as many blocks
645 * since all the data will be copied.
654 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
655 for (s = 0; s < generations[g].n_steps; s++) {
656 if (g == 0 && s == 0) { continue; }
657 stp = &generations[g].steps[s];
658 if (generations[g].steps[0].n_blocks > generations[g].max_blocks) {
659 needed += 2 * stp->n_blocks;
661 needed += stp->n_blocks;
668 /* -----------------------------------------------------------------------------
671 memInventory() checks for memory leaks by counting up all the
672 blocks we know about and comparing that to the number of blocks
673 allegedly floating around in the system.
674 -------------------------------------------------------------------------- */
684 lnat total_blocks = 0, free_blocks = 0;
686 /* count the blocks we current have */
688 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
689 for (s = 0; s < generations[g].n_steps; s++) {
690 stp = &generations[g].steps[s];
691 total_blocks += stp->n_blocks;
692 if (RtsFlags.GcFlags.generations == 1) {
693 /* two-space collector has a to-space too :-) */
694 total_blocks += g0s0->to_blocks;
696 for (bd = stp->large_objects; bd; bd = bd->link) {
697 total_blocks += bd->blocks;
698 /* hack for megablock groups: they have an extra block or two in
699 the second and subsequent megablocks where the block
700 descriptors would normally go.
702 if (bd->blocks > BLOCKS_PER_MBLOCK) {
703 total_blocks -= (MBLOCK_SIZE / BLOCK_SIZE - BLOCKS_PER_MBLOCK)
704 * (bd->blocks/(MBLOCK_SIZE/BLOCK_SIZE));
710 /* any blocks held by allocate() */
711 for (bd = small_alloc_list; bd; bd = bd->link) {
712 total_blocks += bd->blocks;
714 for (bd = large_alloc_list; bd; bd = bd->link) {
715 total_blocks += bd->blocks;
718 /* count the blocks on the free list */
719 free_blocks = countFreeList();
721 ASSERT(total_blocks + free_blocks == mblocks_allocated * BLOCKS_PER_MBLOCK);
724 if (total_blocks + free_blocks != mblocks_allocated *
726 fprintf(stderr, "Blocks: %ld live + %ld free = %ld total (%ld around)\n",
727 total_blocks, free_blocks, total_blocks + free_blocks,
728 mblocks_allocated * BLOCKS_PER_MBLOCK);
733 /* Full heap sanity check. */
740 if (RtsFlags.GcFlags.generations == 1) {
741 checkHeap(g0s0->to_space, NULL);
742 checkChain(g0s0->large_objects);
745 for (g = 0; g <= N; g++) {
746 for (s = 0; s < generations[g].n_steps; s++) {
747 if (g == 0 && s == 0) { continue; }
748 checkHeap(generations[g].steps[s].blocks, NULL);
751 for (g = N+1; g < RtsFlags.GcFlags.generations; g++) {
752 for (s = 0; s < generations[g].n_steps; s++) {
753 checkHeap(generations[g].steps[s].blocks,
754 generations[g].steps[s].blocks->start);
755 checkChain(generations[g].steps[s].large_objects);
758 checkFreeListSanity();