1 /* -----------------------------------------------------------------------------
3 * (c) The GHC Team, 1998-2008
5 * Storage manager front end
7 * Documentation on the architecture of the Storage Manager can be
8 * found in the online commentary:
10 * http://hackage.haskell.org/trac/ghc/wiki/Commentary/Rts/Storage
12 * ---------------------------------------------------------------------------*/
14 #include "PosixSource.h"
20 #include "BlockAlloc.h"
25 #include "OSThreads.h"
26 #include "Capability.h"
29 #include "RetainerProfile.h" // for counting memory blocks (memInventory)
41 * All these globals require sm_mutex to access in THREADED_RTS mode.
43 StgClosure *caf_list = NULL;
44 StgClosure *revertible_caf_list = NULL;
47 bdescr *pinned_object_block; /* allocate pinned objects into this block */
48 nat alloc_blocks; /* number of allocate()d blocks since GC */
49 nat alloc_blocks_lim; /* approximate limit on alloc_blocks */
51 static bdescr *exec_block;
53 generation *generations = NULL; /* all the generations */
54 generation *g0 = NULL; /* generation 0, for convenience */
55 generation *oldest_gen = NULL; /* oldest generation, for convenience */
56 step *g0s0 = NULL; /* generation 0, step 0, for convenience */
59 step *all_steps = NULL; /* single array of steps */
61 ullong total_allocated = 0; /* total memory allocated during run */
63 nat n_nurseries = 0; /* == RtsFlags.ParFlags.nNodes, convenience */
64 step *nurseries = NULL; /* array of nurseries, >1 only if THREADED_RTS */
68 * Storage manager mutex: protects all the above state from
69 * simultaneous access by two STG threads.
73 * This mutex is used by atomicModifyMutVar# only
75 Mutex atomic_modify_mutvar_mutex;
82 static void *stgAllocForGMP (size_t size_in_bytes);
83 static void *stgReallocForGMP (void *ptr, size_t old_size, size_t new_size);
84 static void stgDeallocForGMP (void *ptr, size_t size);
87 initStep (step *stp, int g, int s)
90 stp->abs_no = RtsFlags.GcFlags.steps * g + s;
94 stp->live_estimate = 0;
95 stp->old_blocks = NULL;
96 stp->n_old_blocks = 0;
97 stp->gen = &generations[g];
99 stp->large_objects = NULL;
100 stp->n_large_blocks = 0;
101 stp->scavenged_large_objects = NULL;
102 stp->n_scavenged_large_blocks = 0;
107 initSpinLock(&stp->sync_todo);
108 initSpinLock(&stp->sync_large_objects);
110 stp->threads = END_TSO_QUEUE;
111 stp->old_threads = END_TSO_QUEUE;
120 if (generations != NULL) {
121 // multi-init protection
127 /* Sanity check to make sure the LOOKS_LIKE_ macros appear to be
128 * doing something reasonable.
130 /* We use the NOT_NULL variant or gcc warns that the test is always true */
131 ASSERT(LOOKS_LIKE_INFO_PTR_NOT_NULL((StgWord)&stg_BLACKHOLE_info));
132 ASSERT(LOOKS_LIKE_CLOSURE_PTR(&stg_dummy_ret_closure));
133 ASSERT(!HEAP_ALLOCED(&stg_dummy_ret_closure));
135 if (RtsFlags.GcFlags.maxHeapSize != 0 &&
136 RtsFlags.GcFlags.heapSizeSuggestion >
137 RtsFlags.GcFlags.maxHeapSize) {
138 RtsFlags.GcFlags.maxHeapSize = RtsFlags.GcFlags.heapSizeSuggestion;
141 if (RtsFlags.GcFlags.maxHeapSize != 0 &&
142 RtsFlags.GcFlags.minAllocAreaSize >
143 RtsFlags.GcFlags.maxHeapSize) {
144 errorBelch("maximum heap size (-M) is smaller than minimum alloc area size (-A)");
145 RtsFlags.GcFlags.minAllocAreaSize = RtsFlags.GcFlags.maxHeapSize;
148 initBlockAllocator();
150 #if defined(THREADED_RTS)
151 initMutex(&sm_mutex);
152 initMutex(&atomic_modify_mutvar_mutex);
157 /* allocate generation info array */
158 generations = (generation *)stgMallocBytes(RtsFlags.GcFlags.generations
159 * sizeof(struct generation_),
160 "initStorage: gens");
162 /* allocate all the steps into an array. It is important that we do
163 it this way, because we need the invariant that two step pointers
164 can be directly compared to see which is the oldest.
165 Remember that the last generation has only one step. */
166 total_steps = 1 + (RtsFlags.GcFlags.generations - 1) * RtsFlags.GcFlags.steps;
167 all_steps = stgMallocBytes(total_steps * sizeof(struct step_),
168 "initStorage: steps");
170 /* Initialise all generations */
171 for(g = 0; g < RtsFlags.GcFlags.generations; g++) {
172 gen = &generations[g];
174 gen->mut_list = allocBlock();
175 gen->collections = 0;
176 gen->par_collections = 0;
177 gen->failed_promotions = 0;
181 /* A couple of convenience pointers */
182 g0 = &generations[0];
183 oldest_gen = &generations[RtsFlags.GcFlags.generations-1];
185 /* Allocate step structures in each generation */
186 if (RtsFlags.GcFlags.generations > 1) {
187 /* Only for multiple-generations */
189 /* Oldest generation: one step */
190 oldest_gen->n_steps = 1;
191 oldest_gen->steps = all_steps + (RtsFlags.GcFlags.generations - 1)
192 * RtsFlags.GcFlags.steps;
194 /* set up all except the oldest generation with 2 steps */
195 for(g = 0; g < RtsFlags.GcFlags.generations-1; g++) {
196 generations[g].n_steps = RtsFlags.GcFlags.steps;
197 generations[g].steps = all_steps + g * RtsFlags.GcFlags.steps;
201 /* single generation, i.e. a two-space collector */
203 g0->steps = all_steps;
207 n_nurseries = n_capabilities;
211 nurseries = stgMallocBytes (n_nurseries * sizeof(struct step_),
212 "initStorage: nurseries");
214 /* Initialise all steps */
215 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
216 for (s = 0; s < generations[g].n_steps; s++) {
217 initStep(&generations[g].steps[s], g, s);
221 for (s = 0; s < n_nurseries; s++) {
222 initStep(&nurseries[s], 0, s);
225 /* Set up the destination pointers in each younger gen. step */
226 for (g = 0; g < RtsFlags.GcFlags.generations-1; g++) {
227 for (s = 0; s < generations[g].n_steps-1; s++) {
228 generations[g].steps[s].to = &generations[g].steps[s+1];
230 generations[g].steps[s].to = &generations[g+1].steps[0];
232 oldest_gen->steps[0].to = &oldest_gen->steps[0];
234 for (s = 0; s < n_nurseries; s++) {
235 nurseries[s].to = generations[0].steps[0].to;
238 /* The oldest generation has one step. */
239 if (RtsFlags.GcFlags.compact || RtsFlags.GcFlags.sweep) {
240 if (RtsFlags.GcFlags.generations == 1) {
241 errorBelch("WARNING: compact/sweep is incompatible with -G1; disabled");
243 oldest_gen->steps[0].mark = 1;
244 if (RtsFlags.GcFlags.compact)
245 oldest_gen->steps[0].compact = 1;
249 generations[0].max_blocks = 0;
250 g0s0 = &generations[0].steps[0];
252 /* The allocation area. Policy: keep the allocation area
253 * small to begin with, even if we have a large suggested heap
254 * size. Reason: we're going to do a major collection first, and we
255 * don't want it to be a big one. This vague idea is borne out by
256 * rigorous experimental evidence.
260 weak_ptr_list = NULL;
262 revertible_caf_list = NULL;
264 /* initialise the allocate() interface */
266 alloc_blocks_lim = RtsFlags.GcFlags.minAllocAreaSize;
270 /* Tell GNU multi-precision pkg about our custom alloc functions */
271 mp_set_memory_functions(stgAllocForGMP, stgReallocForGMP, stgDeallocForGMP);
274 initSpinLock(&gc_alloc_block_sync);
275 initSpinLock(&recordMutableGen_sync);
279 IF_DEBUG(gc, statDescribeGens());
287 stat_exit(calcAllocated());
293 stgFree(g0s0); // frees all the steps
294 stgFree(generations);
296 #if defined(THREADED_RTS)
297 closeMutex(&sm_mutex);
298 closeMutex(&atomic_modify_mutvar_mutex);
303 /* -----------------------------------------------------------------------------
306 The entry code for every CAF does the following:
308 - builds a CAF_BLACKHOLE in the heap
309 - pushes an update frame pointing to the CAF_BLACKHOLE
310 - invokes UPD_CAF(), which:
311 - calls newCaf, below
312 - updates the CAF with a static indirection to the CAF_BLACKHOLE
314 Why do we build a BLACKHOLE in the heap rather than just updating
315 the thunk directly? It's so that we only need one kind of update
316 frame - otherwise we'd need a static version of the update frame too.
318 newCaf() does the following:
320 - it puts the CAF on the oldest generation's mut-once list.
321 This is so that we can treat the CAF as a root when collecting
324 For GHCI, we have additional requirements when dealing with CAFs:
326 - we must *retain* all dynamically-loaded CAFs ever entered,
327 just in case we need them again.
328 - we must be able to *revert* CAFs that have been evaluated, to
329 their pre-evaluated form.
331 To do this, we use an additional CAF list. When newCaf() is
332 called on a dynamically-loaded CAF, we add it to the CAF list
333 instead of the old-generation mutable list, and save away its
334 old info pointer (in caf->saved_info) for later reversion.
336 To revert all the CAFs, we traverse the CAF list and reset the
337 info pointer to caf->saved_info, then throw away the CAF list.
338 (see GC.c:revertCAFs()).
342 -------------------------------------------------------------------------- */
345 newCAF(StgClosure* caf)
352 // If we are in GHCi _and_ we are using dynamic libraries,
353 // then we can't redirect newCAF calls to newDynCAF (see below),
354 // so we make newCAF behave almost like newDynCAF.
355 // The dynamic libraries might be used by both the interpreted
356 // program and GHCi itself, so they must not be reverted.
357 // This also means that in GHCi with dynamic libraries, CAFs are not
358 // garbage collected. If this turns out to be a problem, we could
359 // do another hack here and do an address range test on caf to figure
360 // out whether it is from a dynamic library.
361 ((StgIndStatic *)caf)->saved_info = (StgInfoTable *)caf->header.info;
362 ((StgIndStatic *)caf)->static_link = caf_list;
367 /* Put this CAF on the mutable list for the old generation.
368 * This is a HACK - the IND_STATIC closure doesn't really have
369 * a mut_link field, but we pretend it has - in fact we re-use
370 * the STATIC_LINK field for the time being, because when we
371 * come to do a major GC we won't need the mut_link field
372 * any more and can use it as a STATIC_LINK.
374 ((StgIndStatic *)caf)->saved_info = NULL;
375 recordMutableGen(caf, oldest_gen);
381 // An alternate version of newCaf which is used for dynamically loaded
382 // object code in GHCi. In this case we want to retain *all* CAFs in
383 // the object code, because they might be demanded at any time from an
384 // expression evaluated on the command line.
385 // Also, GHCi might want to revert CAFs, so we add these to the
386 // revertible_caf_list.
388 // The linker hackily arranges that references to newCaf from dynamic
389 // code end up pointing to newDynCAF.
391 newDynCAF(StgClosure *caf)
395 ((StgIndStatic *)caf)->saved_info = (StgInfoTable *)caf->header.info;
396 ((StgIndStatic *)caf)->static_link = revertible_caf_list;
397 revertible_caf_list = caf;
402 /* -----------------------------------------------------------------------------
404 -------------------------------------------------------------------------- */
407 allocNursery (step *stp, bdescr *tail, nat blocks)
412 // Allocate a nursery: we allocate fresh blocks one at a time and
413 // cons them on to the front of the list, not forgetting to update
414 // the back pointer on the tail of the list to point to the new block.
415 for (i=0; i < blocks; i++) {
418 processNursery() in LdvProfile.c assumes that every block group in
419 the nursery contains only a single block. So, if a block group is
420 given multiple blocks, change processNursery() accordingly.
424 // double-link the nursery: we might need to insert blocks
431 bd->free = bd->start;
439 assignNurseriesToCapabilities (void)
444 for (i = 0; i < n_nurseries; i++) {
445 capabilities[i].r.rNursery = &nurseries[i];
446 capabilities[i].r.rCurrentNursery = nurseries[i].blocks;
447 capabilities[i].r.rCurrentAlloc = NULL;
449 #else /* THREADED_RTS */
450 MainCapability.r.rNursery = &nurseries[0];
451 MainCapability.r.rCurrentNursery = nurseries[0].blocks;
452 MainCapability.r.rCurrentAlloc = NULL;
457 allocNurseries( void )
461 for (i = 0; i < n_nurseries; i++) {
462 nurseries[i].blocks =
463 allocNursery(&nurseries[i], NULL,
464 RtsFlags.GcFlags.minAllocAreaSize);
465 nurseries[i].n_blocks = RtsFlags.GcFlags.minAllocAreaSize;
466 nurseries[i].old_blocks = NULL;
467 nurseries[i].n_old_blocks = 0;
469 assignNurseriesToCapabilities();
473 resetNurseries( void )
479 for (i = 0; i < n_nurseries; i++) {
481 for (bd = stp->blocks; bd; bd = bd->link) {
482 bd->free = bd->start;
483 ASSERT(bd->gen_no == 0);
484 ASSERT(bd->step == stp);
485 IF_DEBUG(sanity,memset(bd->start, 0xaa, BLOCK_SIZE));
488 assignNurseriesToCapabilities();
492 countNurseryBlocks (void)
497 for (i = 0; i < n_nurseries; i++) {
498 blocks += nurseries[i].n_blocks;
504 resizeNursery ( step *stp, nat blocks )
509 nursery_blocks = stp->n_blocks;
510 if (nursery_blocks == blocks) return;
512 if (nursery_blocks < blocks) {
513 debugTrace(DEBUG_gc, "increasing size of nursery to %d blocks",
515 stp->blocks = allocNursery(stp, stp->blocks, blocks-nursery_blocks);
520 debugTrace(DEBUG_gc, "decreasing size of nursery to %d blocks",
524 while (nursery_blocks > blocks) {
526 next_bd->u.back = NULL;
527 nursery_blocks -= bd->blocks; // might be a large block
532 // might have gone just under, by freeing a large block, so make
533 // up the difference.
534 if (nursery_blocks < blocks) {
535 stp->blocks = allocNursery(stp, stp->blocks, blocks-nursery_blocks);
539 stp->n_blocks = blocks;
540 ASSERT(countBlocks(stp->blocks) == stp->n_blocks);
544 // Resize each of the nurseries to the specified size.
547 resizeNurseriesFixed (nat blocks)
550 for (i = 0; i < n_nurseries; i++) {
551 resizeNursery(&nurseries[i], blocks);
556 // Resize the nurseries to the total specified size.
559 resizeNurseries (nat blocks)
561 // If there are multiple nurseries, then we just divide the number
562 // of available blocks between them.
563 resizeNurseriesFixed(blocks / n_nurseries);
567 /* -----------------------------------------------------------------------------
568 move_TSO is called to update the TSO structure after it has been
569 moved from one place to another.
570 -------------------------------------------------------------------------- */
573 move_TSO (StgTSO *src, StgTSO *dest)
577 // relocate the stack pointer...
578 diff = (StgPtr)dest - (StgPtr)src; // In *words*
579 dest->sp = (StgPtr)dest->sp + diff;
582 /* -----------------------------------------------------------------------------
583 The allocate() interface
585 allocateInGen() function allocates memory directly into a specific
586 generation. It always succeeds, and returns a chunk of memory n
587 words long. n can be larger than the size of a block if necessary,
588 in which case a contiguous block group will be allocated.
590 allocate(n) is equivalent to allocateInGen(g0).
591 -------------------------------------------------------------------------- */
594 allocateInGen (generation *g, lnat n)
602 TICK_ALLOC_HEAP_NOCTR(n);
607 if (n >= LARGE_OBJECT_THRESHOLD/sizeof(W_))
609 lnat req_blocks = (lnat)BLOCK_ROUND_UP(n*sizeof(W_)) / BLOCK_SIZE;
611 // Attempting to allocate an object larger than maxHeapSize
612 // should definitely be disallowed. (bug #1791)
613 if (RtsFlags.GcFlags.maxHeapSize > 0 &&
614 req_blocks >= RtsFlags.GcFlags.maxHeapSize) {
618 bd = allocGroup(req_blocks);
619 dbl_link_onto(bd, &stp->large_objects);
620 stp->n_large_blocks += bd->blocks; // might be larger than req_blocks
623 bd->flags = BF_LARGE;
624 bd->free = bd->start + n;
629 // small allocation (<LARGE_OBJECT_THRESHOLD) */
631 if (bd == NULL || bd->free + n > bd->start + BLOCK_SIZE_W) {
636 bd->link = stp->blocks;
653 return allocateInGen(g0,n);
657 allocatedBytes( void )
661 allocated = alloc_blocks * BLOCK_SIZE_W;
662 if (pinned_object_block != NULL) {
663 allocated -= (pinned_object_block->start + BLOCK_SIZE_W) -
664 pinned_object_block->free;
670 // split N blocks off the front of the given bdescr, returning the
671 // new block group. We treat the remainder as if it
672 // had been freshly allocated in generation 0.
674 splitLargeBlock (bdescr *bd, nat blocks)
678 // subtract the original number of blocks from the counter first
679 bd->step->n_large_blocks -= bd->blocks;
681 new_bd = splitBlockGroup (bd, blocks);
683 dbl_link_onto(new_bd, &g0s0->large_objects);
684 g0s0->n_large_blocks += new_bd->blocks;
685 new_bd->gen_no = g0s0->no;
687 new_bd->flags = BF_LARGE;
688 new_bd->free = bd->free;
689 ASSERT(new_bd->free <= new_bd->start + new_bd->blocks * BLOCK_SIZE_W);
691 // add the new number of blocks to the counter. Due to the gaps
692 // for block descriptor, new_bd->blocks + bd->blocks might not be
693 // equal to the original bd->blocks, which is why we do it this way.
694 bd->step->n_large_blocks += bd->blocks;
699 /* -----------------------------------------------------------------------------
702 This allocates memory in the current thread - it is intended for
703 use primarily from STG-land where we have a Capability. It is
704 better than allocate() because it doesn't require taking the
705 sm_mutex lock in the common case.
707 Memory is allocated directly from the nursery if possible (but not
708 from the current nursery block, so as not to interfere with
710 -------------------------------------------------------------------------- */
713 allocateLocal (Capability *cap, lnat n)
718 if (n >= LARGE_OBJECT_THRESHOLD/sizeof(W_)) {
719 return allocateInGen(g0,n);
722 /* small allocation (<LARGE_OBJECT_THRESHOLD) */
724 TICK_ALLOC_HEAP_NOCTR(n);
727 bd = cap->r.rCurrentAlloc;
728 if (bd == NULL || bd->free + n > bd->start + BLOCK_SIZE_W) {
730 // The CurrentAlloc block is full, we need to find another
731 // one. First, we try taking the next block from the
733 bd = cap->r.rCurrentNursery->link;
735 if (bd == NULL || bd->free + n > bd->start + BLOCK_SIZE_W) {
736 // The nursery is empty, or the next block is already
737 // full: allocate a fresh block (we can't fail here).
740 cap->r.rNursery->n_blocks++;
743 bd->step = cap->r.rNursery;
745 // NO: alloc_blocks++;
746 // calcAllocated() uses the size of the nursery, and we've
747 // already bumpted nursery->n_blocks above.
749 // we have a block in the nursery: take it and put
750 // it at the *front* of the nursery list, and use it
751 // to allocate() from.
752 cap->r.rCurrentNursery->link = bd->link;
753 if (bd->link != NULL) {
754 bd->link->u.back = cap->r.rCurrentNursery;
757 dbl_link_onto(bd, &cap->r.rNursery->blocks);
758 cap->r.rCurrentAlloc = bd;
759 IF_DEBUG(sanity, checkNurserySanity(cap->r.rNursery));
766 /* ---------------------------------------------------------------------------
767 Allocate a fixed/pinned object.
769 We allocate small pinned objects into a single block, allocating a
770 new block when the current one overflows. The block is chained
771 onto the large_object_list of generation 0 step 0.
773 NOTE: The GC can't in general handle pinned objects. This
774 interface is only safe to use for ByteArrays, which have no
775 pointers and don't require scavenging. It works because the
776 block's descriptor has the BF_LARGE flag set, so the block is
777 treated as a large object and chained onto various lists, rather
778 than the individual objects being copied. However, when it comes
779 to scavenge the block, the GC will only scavenge the first object.
780 The reason is that the GC can't linearly scan a block of pinned
781 objects at the moment (doing so would require using the
782 mostly-copying techniques). But since we're restricting ourselves
783 to pinned ByteArrays, not scavenging is ok.
785 This function is called by newPinnedByteArray# which immediately
786 fills the allocated memory with a MutableByteArray#.
787 ------------------------------------------------------------------------- */
790 allocatePinned( lnat n )
793 bdescr *bd = pinned_object_block;
795 // If the request is for a large object, then allocate()
796 // will give us a pinned object anyway.
797 if (n >= LARGE_OBJECT_THRESHOLD/sizeof(W_)) {
803 TICK_ALLOC_HEAP_NOCTR(n);
806 // we always return 8-byte aligned memory. bd->free must be
807 // 8-byte aligned to begin with, so we just round up n to
808 // the nearest multiple of 8 bytes.
809 if (sizeof(StgWord) == 4) {
813 // If we don't have a block of pinned objects yet, or the current
814 // one isn't large enough to hold the new object, allocate a new one.
815 if (bd == NULL || (bd->free + n) > (bd->start + BLOCK_SIZE_W)) {
816 pinned_object_block = bd = allocBlock();
817 dbl_link_onto(bd, &g0s0->large_objects);
818 g0s0->n_large_blocks++;
821 bd->flags = BF_PINNED | BF_LARGE;
822 bd->free = bd->start;
832 /* -----------------------------------------------------------------------------
834 -------------------------------------------------------------------------- */
837 This is the write barrier for MUT_VARs, a.k.a. IORefs. A
838 MUT_VAR_CLEAN object is not on the mutable list; a MUT_VAR_DIRTY
839 is. When written to, a MUT_VAR_CLEAN turns into a MUT_VAR_DIRTY
840 and is put on the mutable list.
843 dirty_MUT_VAR(StgRegTable *reg, StgClosure *p)
845 Capability *cap = regTableToCapability(reg);
847 if (p->header.info == &stg_MUT_VAR_CLEAN_info) {
848 p->header.info = &stg_MUT_VAR_DIRTY_info;
849 bd = Bdescr((StgPtr)p);
850 if (bd->gen_no > 0) recordMutableCap(p,cap,bd->gen_no);
854 // Setting a TSO's link field with a write barrier.
855 // It is *not* necessary to call this function when
856 // * setting the link field to END_TSO_QUEUE
857 // * putting a TSO on the blackhole_queue
858 // * setting the link field of the currently running TSO, as it
859 // will already be dirty.
861 setTSOLink (Capability *cap, StgTSO *tso, StgTSO *target)
864 if ((tso->flags & (TSO_DIRTY|TSO_LINK_DIRTY)) == 0) {
865 tso->flags |= TSO_LINK_DIRTY;
866 bd = Bdescr((StgPtr)tso);
867 if (bd->gen_no > 0) recordMutableCap((StgClosure*)tso,cap,bd->gen_no);
873 dirty_TSO (Capability *cap, StgTSO *tso)
876 if ((tso->flags & (TSO_DIRTY|TSO_LINK_DIRTY)) == 0) {
877 bd = Bdescr((StgPtr)tso);
878 if (bd->gen_no > 0) recordMutableCap((StgClosure*)tso,cap,bd->gen_no);
880 tso->flags |= TSO_DIRTY;
884 This is the write barrier for MVARs. An MVAR_CLEAN objects is not
885 on the mutable list; a MVAR_DIRTY is. When written to, a
886 MVAR_CLEAN turns into a MVAR_DIRTY and is put on the mutable list.
887 The check for MVAR_CLEAN is inlined at the call site for speed,
888 this really does make a difference on concurrency-heavy benchmarks
889 such as Chaneneos and cheap-concurrency.
892 dirty_MVAR(StgRegTable *reg, StgClosure *p)
894 Capability *cap = regTableToCapability(reg);
896 bd = Bdescr((StgPtr)p);
897 if (bd->gen_no > 0) recordMutableCap(p,cap,bd->gen_no);
900 /* -----------------------------------------------------------------------------
901 Allocation functions for GMP.
903 These all use the allocate() interface - we can't have any garbage
904 collection going on during a gmp operation, so we use allocate()
905 which always succeeds. The gmp operations which might need to
906 allocate will ask the storage manager (via doYouWantToGC()) whether
907 a garbage collection is required, in case we get into a loop doing
908 only allocate() style allocation.
909 -------------------------------------------------------------------------- */
912 stgAllocForGMP (size_t size_in_bytes)
915 nat data_size_in_words, total_size_in_words;
917 /* round up to a whole number of words */
918 data_size_in_words = (size_in_bytes + sizeof(W_) + 1) / sizeof(W_);
919 total_size_in_words = sizeofW(StgArrWords) + data_size_in_words;
921 /* allocate and fill it in. */
922 #if defined(THREADED_RTS)
923 arr = (StgArrWords *)allocateLocal(myTask()->cap, total_size_in_words);
925 arr = (StgArrWords *)allocateLocal(&MainCapability, total_size_in_words);
927 SET_ARR_HDR(arr, &stg_ARR_WORDS_info, CCCS, data_size_in_words);
929 /* and return a ptr to the goods inside the array */
934 stgReallocForGMP (void *ptr, size_t old_size, size_t new_size)
937 void *new_stuff_ptr = stgAllocForGMP(new_size);
939 char *p = (char *) ptr;
940 char *q = (char *) new_stuff_ptr;
942 min_size = old_size < new_size ? old_size : new_size;
943 for (; i < min_size; i++, p++, q++) {
947 return(new_stuff_ptr);
951 stgDeallocForGMP (void *ptr STG_UNUSED,
952 size_t size STG_UNUSED)
954 /* easy for us: the garbage collector does the dealloc'n */
957 /* -----------------------------------------------------------------------------
959 * -------------------------------------------------------------------------- */
961 /* -----------------------------------------------------------------------------
964 * Approximate how much we've allocated: number of blocks in the
965 * nursery + blocks allocated via allocate() - unused nusery blocks.
966 * This leaves a little slop at the end of each block, and doesn't
967 * take into account large objects (ToDo).
968 * -------------------------------------------------------------------------- */
971 calcAllocated( void )
976 allocated = allocatedBytes();
977 allocated += countNurseryBlocks() * BLOCK_SIZE_W;
982 for (i = 0; i < n_nurseries; i++) {
984 for ( bd = capabilities[i].r.rCurrentNursery->link;
985 bd != NULL; bd = bd->link ) {
986 allocated -= BLOCK_SIZE_W;
988 cap = &capabilities[i];
989 if (cap->r.rCurrentNursery->free <
990 cap->r.rCurrentNursery->start + BLOCK_SIZE_W) {
991 allocated -= (cap->r.rCurrentNursery->start + BLOCK_SIZE_W)
992 - cap->r.rCurrentNursery->free;
996 bdescr *current_nursery = MainCapability.r.rCurrentNursery;
998 for ( bd = current_nursery->link; bd != NULL; bd = bd->link ) {
999 allocated -= BLOCK_SIZE_W;
1001 if (current_nursery->free < current_nursery->start + BLOCK_SIZE_W) {
1002 allocated -= (current_nursery->start + BLOCK_SIZE_W)
1003 - current_nursery->free;
1008 total_allocated += allocated;
1012 /* Approximate the amount of live data in the heap. To be called just
1013 * after garbage collection (see GarbageCollect()).
1016 calcLiveBlocks(void)
1022 if (RtsFlags.GcFlags.generations == 1) {
1023 return g0s0->n_large_blocks + g0s0->n_blocks;
1026 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1027 for (s = 0; s < generations[g].n_steps; s++) {
1028 /* approximate amount of live data (doesn't take into account slop
1029 * at end of each block).
1031 if (g == 0 && s == 0) {
1034 stp = &generations[g].steps[s];
1035 live += stp->n_large_blocks + stp->n_blocks;
1042 countOccupied(bdescr *bd)
1047 for (; bd != NULL; bd = bd->link) {
1048 ASSERT(bd->free <= bd->start + bd->blocks * BLOCK_SIZE_W);
1049 words += bd->free - bd->start;
1054 // Return an accurate count of the live data in the heap, excluding
1063 if (RtsFlags.GcFlags.generations == 1) {
1064 return g0s0->n_words + countOccupied(g0s0->large_objects);
1068 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1069 for (s = 0; s < generations[g].n_steps; s++) {
1070 if (g == 0 && s == 0) continue;
1071 stp = &generations[g].steps[s];
1072 live += stp->n_words + countOccupied(stp->large_objects);
1078 /* Approximate the number of blocks that will be needed at the next
1079 * garbage collection.
1081 * Assume: all data currently live will remain live. Steps that will
1082 * be collected next time will therefore need twice as many blocks
1083 * since all the data will be copied.
1092 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1093 for (s = 0; s < generations[g].n_steps; s++) {
1094 if (g == 0 && s == 0) { continue; }
1095 stp = &generations[g].steps[s];
1097 // we need at least this much space
1098 needed += stp->n_blocks + stp->n_large_blocks;
1100 // any additional space needed to collect this gen next time?
1101 if (g == 0 || // always collect gen 0
1102 (generations[g].steps[0].n_blocks +
1103 generations[g].steps[0].n_large_blocks
1104 > generations[g].max_blocks)) {
1105 // we will collect this gen next time
1108 needed += stp->n_blocks / BITS_IN(W_);
1110 needed += stp->n_blocks / 100;
1113 continue; // no additional space needed for compaction
1115 needed += stp->n_blocks;
1123 /* ----------------------------------------------------------------------------
1126 Executable memory must be managed separately from non-executable
1127 memory. Most OSs these days require you to jump through hoops to
1128 dynamically allocate executable memory, due to various security
1131 Here we provide a small memory allocator for executable memory.
1132 Memory is managed with a page granularity; we allocate linearly
1133 in the page, and when the page is emptied (all objects on the page
1134 are free) we free the page again, not forgetting to make it
1137 TODO: The inability to handle objects bigger than BLOCK_SIZE_W means that
1138 the linker cannot use allocateExec for loading object code files
1139 on Windows. Once allocateExec can handle larger objects, the linker
1140 should be modified to use allocateExec instead of VirtualAlloc.
1141 ------------------------------------------------------------------------- */
1143 #if defined(linux_HOST_OS)
1145 // On Linux we need to use libffi for allocating executable memory,
1146 // because it knows how to work around the restrictions put in place
1149 void *allocateExec (nat bytes, void **exec_ret)
1153 ret = ffi_closure_alloc (sizeof(void *) + (size_t)bytes, (void**)&exec);
1155 if (ret == NULL) return ret;
1156 *ret = ret; // save the address of the writable mapping, for freeExec().
1157 *exec_ret = exec + 1;
1161 // freeExec gets passed the executable address, not the writable address.
1162 void freeExec (void *addr)
1165 writable = *((void**)addr - 1);
1167 ffi_closure_free (writable);
1173 void *allocateExec (nat bytes, void **exec_ret)
1180 // round up to words.
1181 n = (bytes + sizeof(W_) + 1) / sizeof(W_);
1183 if (n+1 > BLOCK_SIZE_W) {
1184 barf("allocateExec: can't handle large objects");
1187 if (exec_block == NULL ||
1188 exec_block->free + n + 1 > exec_block->start + BLOCK_SIZE_W) {
1190 lnat pagesize = getPageSize();
1191 bd = allocGroup(stg_max(1, pagesize / BLOCK_SIZE));
1192 debugTrace(DEBUG_gc, "allocate exec block %p", bd->start);
1194 bd->flags = BF_EXEC;
1195 bd->link = exec_block;
1196 if (exec_block != NULL) {
1197 exec_block->u.back = bd;
1200 setExecutable(bd->start, bd->blocks * BLOCK_SIZE, rtsTrue);
1203 *(exec_block->free) = n; // store the size of this chunk
1204 exec_block->gen_no += n; // gen_no stores the number of words allocated
1205 ret = exec_block->free + 1;
1206 exec_block->free += n + 1;
1213 void freeExec (void *addr)
1215 StgPtr p = (StgPtr)addr - 1;
1216 bdescr *bd = Bdescr((StgPtr)p);
1218 if ((bd->flags & BF_EXEC) == 0) {
1219 barf("freeExec: not executable");
1222 if (*(StgPtr)p == 0) {
1223 barf("freeExec: already free?");
1228 bd->gen_no -= *(StgPtr)p;
1231 if (bd->gen_no == 0) {
1232 // Free the block if it is empty, but not if it is the block at
1233 // the head of the queue.
1234 if (bd != exec_block) {
1235 debugTrace(DEBUG_gc, "free exec block %p", bd->start);
1236 dbl_link_remove(bd, &exec_block);
1237 setExecutable(bd->start, bd->blocks * BLOCK_SIZE, rtsFalse);
1240 bd->free = bd->start;
1247 #endif /* mingw32_HOST_OS */
1249 /* -----------------------------------------------------------------------------
1252 memInventory() checks for memory leaks by counting up all the
1253 blocks we know about and comparing that to the number of blocks
1254 allegedly floating around in the system.
1255 -------------------------------------------------------------------------- */
1259 // Useful for finding partially full blocks in gdb
1260 void findSlop(bdescr *bd);
1261 void findSlop(bdescr *bd)
1265 for (; bd != NULL; bd = bd->link) {
1266 slop = (bd->blocks * BLOCK_SIZE_W) - (bd->free - bd->start);
1267 if (slop > (1024/sizeof(W_))) {
1268 debugBelch("block at %p (bdescr %p) has %ldKB slop\n",
1269 bd->start, bd, slop / (1024/sizeof(W_)));
1275 countBlocks(bdescr *bd)
1278 for (n=0; bd != NULL; bd=bd->link) {
1284 // (*1) Just like countBlocks, except that we adjust the count for a
1285 // megablock group so that it doesn't include the extra few blocks
1286 // that would be taken up by block descriptors in the second and
1287 // subsequent megablock. This is so we can tally the count with the
1288 // number of blocks allocated in the system, for memInventory().
1290 countAllocdBlocks(bdescr *bd)
1293 for (n=0; bd != NULL; bd=bd->link) {
1295 // hack for megablock groups: see (*1) above
1296 if (bd->blocks > BLOCKS_PER_MBLOCK) {
1297 n -= (MBLOCK_SIZE / BLOCK_SIZE - BLOCKS_PER_MBLOCK)
1298 * (bd->blocks/(MBLOCK_SIZE/BLOCK_SIZE));
1305 stepBlocks (step *stp)
1307 ASSERT(countBlocks(stp->blocks) == stp->n_blocks);
1308 ASSERT(countBlocks(stp->large_objects) == stp->n_large_blocks);
1309 return stp->n_blocks + stp->n_old_blocks +
1310 countAllocdBlocks(stp->large_objects);
1313 // If memInventory() calculates that we have a memory leak, this
1314 // function will try to find the block(s) that are leaking by marking
1315 // all the ones that we know about, and search through memory to find
1316 // blocks that are not marked. In the debugger this can help to give
1317 // us a clue about what kind of block leaked. In the future we might
1318 // annotate blocks with their allocation site to give more helpful
1321 findMemoryLeak (void)
1324 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1325 for (i = 0; i < n_capabilities; i++) {
1326 markBlocks(capabilities[i].mut_lists[g]);
1328 markBlocks(generations[g].mut_list);
1329 for (s = 0; s < generations[g].n_steps; s++) {
1330 markBlocks(generations[g].steps[s].blocks);
1331 markBlocks(generations[g].steps[s].large_objects);
1335 for (i = 0; i < n_nurseries; i++) {
1336 markBlocks(nurseries[i].blocks);
1337 markBlocks(nurseries[i].large_objects);
1342 // if (RtsFlags.ProfFlags.doHeapProfile == HEAP_BY_RETAINER) {
1343 // markRetainerBlocks();
1347 // count the blocks allocated by the arena allocator
1349 // markArenaBlocks();
1351 // count the blocks containing executable memory
1352 markBlocks(exec_block);
1354 reportUnmarkedBlocks();
1359 memInventory (rtsBool show)
1363 lnat gen_blocks[RtsFlags.GcFlags.generations];
1364 lnat nursery_blocks, retainer_blocks,
1365 arena_blocks, exec_blocks;
1366 lnat live_blocks = 0, free_blocks = 0;
1369 // count the blocks we current have
1371 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1373 for (i = 0; i < n_capabilities; i++) {
1374 gen_blocks[g] += countBlocks(capabilities[i].mut_lists[g]);
1376 gen_blocks[g] += countAllocdBlocks(generations[g].mut_list);
1377 for (s = 0; s < generations[g].n_steps; s++) {
1378 stp = &generations[g].steps[s];
1379 gen_blocks[g] += stepBlocks(stp);
1384 for (i = 0; i < n_nurseries; i++) {
1385 nursery_blocks += stepBlocks(&nurseries[i]);
1388 retainer_blocks = 0;
1390 if (RtsFlags.ProfFlags.doHeapProfile == HEAP_BY_RETAINER) {
1391 retainer_blocks = retainerStackBlocks();
1395 // count the blocks allocated by the arena allocator
1396 arena_blocks = arenaBlocks();
1398 // count the blocks containing executable memory
1399 exec_blocks = countAllocdBlocks(exec_block);
1401 /* count the blocks on the free list */
1402 free_blocks = countFreeList();
1405 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1406 live_blocks += gen_blocks[g];
1408 live_blocks += nursery_blocks +
1409 + retainer_blocks + arena_blocks + exec_blocks;
1411 #define MB(n) (((n) * BLOCK_SIZE_W) / ((1024*1024)/sizeof(W_)))
1413 leak = live_blocks + free_blocks != mblocks_allocated * BLOCKS_PER_MBLOCK;
1418 debugBelch("Memory leak detected:\n");
1420 debugBelch("Memory inventory:\n");
1422 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1423 debugBelch(" gen %d blocks : %5lu blocks (%lu MB)\n", g,
1424 gen_blocks[g], MB(gen_blocks[g]));
1426 debugBelch(" nursery : %5lu blocks (%lu MB)\n",
1427 nursery_blocks, MB(nursery_blocks));
1428 debugBelch(" retainer : %5lu blocks (%lu MB)\n",
1429 retainer_blocks, MB(retainer_blocks));
1430 debugBelch(" arena blocks : %5lu blocks (%lu MB)\n",
1431 arena_blocks, MB(arena_blocks));
1432 debugBelch(" exec : %5lu blocks (%lu MB)\n",
1433 exec_blocks, MB(exec_blocks));
1434 debugBelch(" free : %5lu blocks (%lu MB)\n",
1435 free_blocks, MB(free_blocks));
1436 debugBelch(" total : %5lu blocks (%lu MB)\n",
1437 live_blocks + free_blocks, MB(live_blocks+free_blocks));
1439 debugBelch("\n in system : %5lu blocks (%lu MB)\n",
1440 mblocks_allocated * BLOCKS_PER_MBLOCK, mblocks_allocated);
1448 ASSERT(n_alloc_blocks == live_blocks);
1453 /* Full heap sanity check. */
1459 if (RtsFlags.GcFlags.generations == 1) {
1460 checkHeap(g0s0->blocks);
1461 checkChain(g0s0->large_objects);
1464 for (g = 0; g < RtsFlags.GcFlags.generations; g++) {
1465 for (s = 0; s < generations[g].n_steps; s++) {
1466 if (g == 0 && s == 0) { continue; }
1467 ASSERT(countBlocks(generations[g].steps[s].blocks)
1468 == generations[g].steps[s].n_blocks);
1469 ASSERT(countBlocks(generations[g].steps[s].large_objects)
1470 == generations[g].steps[s].n_large_blocks);
1471 checkHeap(generations[g].steps[s].blocks);
1472 checkChain(generations[g].steps[s].large_objects);
1474 checkMutableList(generations[g].mut_list, g);
1479 for (s = 0; s < n_nurseries; s++) {
1480 ASSERT(countBlocks(nurseries[s].blocks)
1481 == nurseries[s].n_blocks);
1482 ASSERT(countBlocks(nurseries[s].large_objects)
1483 == nurseries[s].n_large_blocks);
1486 checkFreeListSanity();
1489 #if defined(THREADED_RTS)
1490 // check the stacks too in threaded mode, because we don't do a
1491 // full heap sanity check in this case (see checkHeap())
1492 checkGlobalTSOList(rtsTrue);
1494 checkGlobalTSOList(rtsFalse);
1498 /* Nursery sanity check */
1500 checkNurserySanity( step *stp )
1506 for (bd = stp->blocks; bd != NULL; bd = bd->link) {
1507 ASSERT(bd->u.back == prev);
1509 blocks += bd->blocks;
1511 ASSERT(blocks == stp->n_blocks);
1514 // handy function for use in gdb, because Bdescr() is inlined.
1515 extern bdescr *_bdescr( StgPtr p );