/* * Copyright (c) 2001, 2011, Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com if you need additional information or have any * questions. * */ #include "precompiled.hpp" #include "gc_implementation/g1/concurrentG1Refine.hpp" #include "gc_implementation/g1/g1BlockOffsetTable.inline.hpp" #include "gc_implementation/g1/g1CollectedHeap.inline.hpp" #include "gc_implementation/g1/heapRegionRemSet.hpp" #include "gc_implementation/g1/heapRegionSeq.inline.hpp" #include "memory/allocation.hpp" #include "memory/space.inline.hpp" #include "utilities/bitMap.inline.hpp" #include "utilities/globalDefinitions.hpp" #define HRRS_VERBOSE 0 #define PRT_COUNT_OCCUPIED 1 // OtherRegionsTable class PerRegionTable: public CHeapObj { friend class OtherRegionsTable; friend class HeapRegionRemSetIterator; HeapRegion* _hr; BitMap _bm; #if PRT_COUNT_OCCUPIED jint _occupied; #endif PerRegionTable* _next_free; PerRegionTable* next_free() { return _next_free; } void set_next_free(PerRegionTable* prt) { _next_free = prt; } static PerRegionTable* _free_list; #ifdef _MSC_VER // For some reason even though the classes are marked as friend they are unable // to access CardsPerRegion when private/protected. Only the windows c++ compiler // says this Sun CC and linux gcc don't have a problem with access when private public: #endif // _MSC_VER protected: // We need access in order to union things into the base table. BitMap* bm() { return &_bm; } #if PRT_COUNT_OCCUPIED void recount_occupied() { _occupied = (jint) bm()->count_one_bits(); } #endif PerRegionTable(HeapRegion* hr) : _hr(hr), #if PRT_COUNT_OCCUPIED _occupied(0), #endif _bm(HeapRegion::CardsPerRegion, false /* in-resource-area */) {} static void free(PerRegionTable* prt) { while (true) { PerRegionTable* fl = _free_list; prt->set_next_free(fl); PerRegionTable* res = (PerRegionTable*) Atomic::cmpxchg_ptr(prt, &_free_list, fl); if (res == fl) return; } ShouldNotReachHere(); } static PerRegionTable* alloc(HeapRegion* hr) { PerRegionTable* fl = _free_list; while (fl != NULL) { PerRegionTable* nxt = fl->next_free(); PerRegionTable* res = (PerRegionTable*) Atomic::cmpxchg_ptr(nxt, &_free_list, fl); if (res == fl) { fl->init(hr); return fl; } else { fl = _free_list; } } assert(fl == NULL, "Loop condition."); return new PerRegionTable(hr); } void add_card_work(CardIdx_t from_card, bool par) { if (!_bm.at(from_card)) { if (par) { if (_bm.par_at_put(from_card, 1)) { #if PRT_COUNT_OCCUPIED Atomic::inc(&_occupied); #endif } } else { _bm.at_put(from_card, 1); #if PRT_COUNT_OCCUPIED _occupied++; #endif } } } void add_reference_work(OopOrNarrowOopStar from, bool par) { // Must make this robust in case "from" is not in "_hr", because of // concurrency. #if HRRS_VERBOSE gclog_or_tty->print_cr(" PRT::Add_reference_work(" PTR_FORMAT "->" PTR_FORMAT").", from, *from); #endif HeapRegion* loc_hr = hr(); // If the test below fails, then this table was reused concurrently // with this operation. This is OK, since the old table was coarsened, // and adding a bit to the new table is never incorrect. if (loc_hr->is_in_reserved_raw(from)) { size_t hw_offset = pointer_delta((HeapWord*)from, loc_hr->bottom()); CardIdx_t from_card = (CardIdx_t) hw_offset >> (CardTableModRefBS::card_shift - LogHeapWordSize); assert(0 <= from_card && (size_t)from_card < HeapRegion::CardsPerRegion, "Must be in range."); add_card_work(from_card, par); } } public: HeapRegion* hr() const { return _hr; } #if PRT_COUNT_OCCUPIED jint occupied() const { // Overkill, but if we ever need it... // guarantee(_occupied == _bm.count_one_bits(), "Check"); return _occupied; } #else jint occupied() const { return _bm.count_one_bits(); } #endif void init(HeapRegion* hr) { _hr = hr; #if PRT_COUNT_OCCUPIED _occupied = 0; #endif _bm.clear(); } void add_reference(OopOrNarrowOopStar from) { add_reference_work(from, /*parallel*/ true); } void seq_add_reference(OopOrNarrowOopStar from) { add_reference_work(from, /*parallel*/ false); } void scrub(CardTableModRefBS* ctbs, BitMap* card_bm) { HeapWord* hr_bot = hr()->bottom(); size_t hr_first_card_index = ctbs->index_for(hr_bot); bm()->set_intersection_at_offset(*card_bm, hr_first_card_index); #if PRT_COUNT_OCCUPIED recount_occupied(); #endif } void add_card(CardIdx_t from_card_index) { add_card_work(from_card_index, /*parallel*/ true); } void seq_add_card(CardIdx_t from_card_index) { add_card_work(from_card_index, /*parallel*/ false); } // (Destructively) union the bitmap of the current table into the given // bitmap (which is assumed to be of the same size.) void union_bitmap_into(BitMap* bm) { bm->set_union(_bm); } // Mem size in bytes. size_t mem_size() const { return sizeof(this) + _bm.size_in_words() * HeapWordSize; } static size_t fl_mem_size() { PerRegionTable* cur = _free_list; size_t res = 0; while (cur != NULL) { res += sizeof(PerRegionTable); cur = cur->next_free(); } return res; } // Requires "from" to be in "hr()". bool contains_reference(OopOrNarrowOopStar from) const { assert(hr()->is_in_reserved(from), "Precondition."); size_t card_ind = pointer_delta(from, hr()->bottom(), CardTableModRefBS::card_size); return _bm.at(card_ind); } }; PerRegionTable* PerRegionTable::_free_list = NULL; #define COUNT_PAR_EXPANDS 0 #if COUNT_PAR_EXPANDS static jint n_par_expands = 0; static jint n_par_contracts = 0; static jint par_expand_list_len = 0; static jint max_par_expand_list_len = 0; static void print_par_expand() { Atomic::inc(&n_par_expands); Atomic::inc(&par_expand_list_len); if (par_expand_list_len > max_par_expand_list_len) { max_par_expand_list_len = par_expand_list_len; } if ((n_par_expands % 10) == 0) { gclog_or_tty->print_cr("\n\n%d par expands: %d contracts, " "len = %d, max_len = %d\n.", n_par_expands, n_par_contracts, par_expand_list_len, max_par_expand_list_len); } } #endif class PosParPRT: public PerRegionTable { PerRegionTable** _par_tables; enum SomePrivateConstants { ReserveParTableExpansion = 1 }; void par_contract() { assert(_par_tables != NULL, "Precondition."); int n = HeapRegionRemSet::num_par_rem_sets()-1; for (int i = 0; i < n; i++) { _par_tables[i]->union_bitmap_into(bm()); PerRegionTable::free(_par_tables[i]); _par_tables[i] = NULL; } #if PRT_COUNT_OCCUPIED // We must recount the "occupied." recount_occupied(); #endif FREE_C_HEAP_ARRAY(PerRegionTable*, _par_tables); _par_tables = NULL; #if COUNT_PAR_EXPANDS Atomic::inc(&n_par_contracts); Atomic::dec(&par_expand_list_len); #endif } static PerRegionTable** _par_table_fl; PosParPRT* _next; static PosParPRT* _free_list; PerRegionTable** par_tables() const { assert(uintptr_t(NULL) == 0, "Assumption."); if (uintptr_t(_par_tables) <= ReserveParTableExpansion) return NULL; else return _par_tables; } PosParPRT* _next_par_expanded; PosParPRT* next_par_expanded() { return _next_par_expanded; } void set_next_par_expanded(PosParPRT* ppprt) { _next_par_expanded = ppprt; } static PosParPRT* _par_expanded_list; public: PosParPRT(HeapRegion* hr) : PerRegionTable(hr), _par_tables(NULL) {} jint occupied() const { jint res = PerRegionTable::occupied(); if (par_tables() != NULL) { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets()-1; i++) { res += par_tables()[i]->occupied(); } } return res; } void init(HeapRegion* hr) { PerRegionTable::init(hr); _next = NULL; if (par_tables() != NULL) { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets()-1; i++) { par_tables()[i]->init(hr); } } } static void free(PosParPRT* prt) { while (true) { PosParPRT* fl = _free_list; prt->set_next(fl); PosParPRT* res = (PosParPRT*) Atomic::cmpxchg_ptr(prt, &_free_list, fl); if (res == fl) return; } ShouldNotReachHere(); } static PosParPRT* alloc(HeapRegion* hr) { PosParPRT* fl = _free_list; while (fl != NULL) { PosParPRT* nxt = fl->next(); PosParPRT* res = (PosParPRT*) Atomic::cmpxchg_ptr(nxt, &_free_list, fl); if (res == fl) { fl->init(hr); return fl; } else { fl = _free_list; } } assert(fl == NULL, "Loop condition."); return new PosParPRT(hr); } PosParPRT* next() const { return _next; } void set_next(PosParPRT* nxt) { _next = nxt; } PosParPRT** next_addr() { return &_next; } bool should_expand(int tid) { // Given that we now defer RSet updates for after a GC we don't // really need to expand the tables any more. This code should be // cleaned up in the future (see CR 6921087). return false; } void par_expand() { int n = HeapRegionRemSet::num_par_rem_sets()-1; if (n <= 0) return; if (_par_tables == NULL) { PerRegionTable* res = (PerRegionTable*) Atomic::cmpxchg_ptr((PerRegionTable*)ReserveParTableExpansion, &_par_tables, NULL); if (res != NULL) return; // Otherwise, we reserved the right to do the expansion. PerRegionTable** ptables = NEW_C_HEAP_ARRAY(PerRegionTable*, n); for (int i = 0; i < n; i++) { PerRegionTable* ptable = PerRegionTable::alloc(hr()); ptables[i] = ptable; } // Here we do not need an atomic. _par_tables = ptables; #if COUNT_PAR_EXPANDS print_par_expand(); #endif // We must put this table on the expanded list. PosParPRT* exp_head = _par_expanded_list; while (true) { set_next_par_expanded(exp_head); PosParPRT* res = (PosParPRT*) Atomic::cmpxchg_ptr(this, &_par_expanded_list, exp_head); if (res == exp_head) return; // Otherwise. exp_head = res; } ShouldNotReachHere(); } } void add_reference(OopOrNarrowOopStar from, int tid) { // Expand if necessary. PerRegionTable** pt = par_tables(); if (pt != NULL) { // We always have to assume that mods to table 0 are in parallel, // because of the claiming scheme in parallel expansion. A thread // with tid != 0 that finds the table to be NULL, but doesn't succeed // in claiming the right of expanding it, will end up in the else // clause of the above if test. That thread could be delayed, and a // thread 0 add reference could see the table expanded, and come // here. Both threads would be adding in parallel. But we get to // not use atomics for tids > 0. if (tid == 0) { PerRegionTable::add_reference(from); } else { pt[tid-1]->seq_add_reference(from); } } else { // Not expanded -- add to the base table. PerRegionTable::add_reference(from); } } void scrub(CardTableModRefBS* ctbs, BitMap* card_bm) { assert(_par_tables == NULL, "Precondition"); PerRegionTable::scrub(ctbs, card_bm); } size_t mem_size() const { size_t res = PerRegionTable::mem_size() + sizeof(this) - sizeof(PerRegionTable); if (_par_tables != NULL) { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets()-1; i++) { res += _par_tables[i]->mem_size(); } } return res; } static size_t fl_mem_size() { PosParPRT* cur = _free_list; size_t res = 0; while (cur != NULL) { res += sizeof(PosParPRT); cur = cur->next(); } return res; } bool contains_reference(OopOrNarrowOopStar from) const { if (PerRegionTable::contains_reference(from)) return true; if (_par_tables != NULL) { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets()-1; i++) { if (_par_tables[i]->contains_reference(from)) return true; } } return false; } static void par_contract_all(); }; void PosParPRT::par_contract_all() { PosParPRT* hd = _par_expanded_list; while (hd != NULL) { PosParPRT* nxt = hd->next_par_expanded(); PosParPRT* res = (PosParPRT*) Atomic::cmpxchg_ptr(nxt, &_par_expanded_list, hd); if (res == hd) { // We claimed the right to contract this table. hd->set_next_par_expanded(NULL); hd->par_contract(); hd = _par_expanded_list; } else { hd = res; } } } PosParPRT* PosParPRT::_free_list = NULL; PosParPRT* PosParPRT::_par_expanded_list = NULL; jint OtherRegionsTable::_cache_probes = 0; jint OtherRegionsTable::_cache_hits = 0; size_t OtherRegionsTable::_max_fine_entries = 0; size_t OtherRegionsTable::_mod_max_fine_entries_mask = 0; #if SAMPLE_FOR_EVICTION size_t OtherRegionsTable::_fine_eviction_stride = 0; size_t OtherRegionsTable::_fine_eviction_sample_size = 0; #endif OtherRegionsTable::OtherRegionsTable(HeapRegion* hr) : _g1h(G1CollectedHeap::heap()), _m(Mutex::leaf, "An OtherRegionsTable lock", true), _hr(hr), _coarse_map(G1CollectedHeap::heap()->max_regions(), false /* in-resource-area */), _fine_grain_regions(NULL), _n_fine_entries(0), _n_coarse_entries(0), #if SAMPLE_FOR_EVICTION _fine_eviction_start(0), #endif _sparse_table(hr) { typedef PosParPRT* PosParPRTPtr; if (_max_fine_entries == 0) { assert(_mod_max_fine_entries_mask == 0, "Both or none."); size_t max_entries_log = (size_t)log2_long((jlong)G1RSetRegionEntries); _max_fine_entries = (size_t)(1 << max_entries_log); _mod_max_fine_entries_mask = _max_fine_entries - 1; #if SAMPLE_FOR_EVICTION assert(_fine_eviction_sample_size == 0 && _fine_eviction_stride == 0, "All init at same time."); _fine_eviction_sample_size = MAX2((size_t)4, max_entries_log); _fine_eviction_stride = _max_fine_entries / _fine_eviction_sample_size; #endif } _fine_grain_regions = new PosParPRTPtr[_max_fine_entries]; if (_fine_grain_regions == NULL) vm_exit_out_of_memory(sizeof(void*)*_max_fine_entries, "Failed to allocate _fine_grain_entries."); for (size_t i = 0; i < _max_fine_entries; i++) { _fine_grain_regions[i] = NULL; } } int** OtherRegionsTable::_from_card_cache = NULL; size_t OtherRegionsTable::_from_card_cache_max_regions = 0; size_t OtherRegionsTable::_from_card_cache_mem_size = 0; void OtherRegionsTable::init_from_card_cache(size_t max_regions) { _from_card_cache_max_regions = max_regions; int n_par_rs = HeapRegionRemSet::num_par_rem_sets(); _from_card_cache = NEW_C_HEAP_ARRAY(int*, n_par_rs); for (int i = 0; i < n_par_rs; i++) { _from_card_cache[i] = NEW_C_HEAP_ARRAY(int, max_regions); for (size_t j = 0; j < max_regions; j++) { _from_card_cache[i][j] = -1; // An invalid value. } } _from_card_cache_mem_size = n_par_rs * max_regions * sizeof(int); } void OtherRegionsTable::shrink_from_card_cache(size_t new_n_regs) { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets(); i++) { assert(new_n_regs <= _from_card_cache_max_regions, "Must be within max."); for (size_t j = new_n_regs; j < _from_card_cache_max_regions; j++) { _from_card_cache[i][j] = -1; // An invalid value. } } } #ifndef PRODUCT void OtherRegionsTable::print_from_card_cache() { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets(); i++) { for (size_t j = 0; j < _from_card_cache_max_regions; j++) { gclog_or_tty->print_cr("_from_card_cache[%d][%d] = %d.", i, j, _from_card_cache[i][j]); } } } #endif void OtherRegionsTable::add_reference(OopOrNarrowOopStar from, int tid) { size_t cur_hrs_ind = hr()->hrs_index(); #if HRRS_VERBOSE gclog_or_tty->print_cr("ORT::add_reference_work(" PTR_FORMAT "->" PTR_FORMAT ").", from, UseCompressedOops ? oopDesc::load_decode_heap_oop((narrowOop*)from) : oopDesc::load_decode_heap_oop((oop*)from)); #endif int from_card = (int)(uintptr_t(from) >> CardTableModRefBS::card_shift); #if HRRS_VERBOSE gclog_or_tty->print_cr("Table for [" PTR_FORMAT "...): card %d (cache = %d)", hr()->bottom(), from_card, _from_card_cache[tid][cur_hrs_ind]); #endif #define COUNT_CACHE 0 #if COUNT_CACHE jint p = Atomic::add(1, &_cache_probes); if ((p % 10000) == 0) { jint hits = _cache_hits; gclog_or_tty->print_cr("%d/%d = %5.2f%% RS cache hits.", _cache_hits, p, 100.0* (float)hits/(float)p); } #endif if (from_card == _from_card_cache[tid][cur_hrs_ind]) { #if HRRS_VERBOSE gclog_or_tty->print_cr(" from-card cache hit."); #endif #if COUNT_CACHE Atomic::inc(&_cache_hits); #endif assert(contains_reference(from), "We just added it!"); return; } else { _from_card_cache[tid][cur_hrs_ind] = from_card; } // Note that this may be a continued H region. HeapRegion* from_hr = _g1h->heap_region_containing_raw(from); RegionIdx_t from_hrs_ind = (RegionIdx_t) from_hr->hrs_index(); // If the region is already coarsened, return. if (_coarse_map.at(from_hrs_ind)) { #if HRRS_VERBOSE gclog_or_tty->print_cr(" coarse map hit."); #endif assert(contains_reference(from), "We just added it!"); return; } // Otherwise find a per-region table to add it to. size_t ind = from_hrs_ind & _mod_max_fine_entries_mask; PosParPRT* prt = find_region_table(ind, from_hr); if (prt == NULL) { MutexLockerEx x(&_m, Mutex::_no_safepoint_check_flag); // Confirm that it's really not there... prt = find_region_table(ind, from_hr); if (prt == NULL) { uintptr_t from_hr_bot_card_index = uintptr_t(from_hr->bottom()) >> CardTableModRefBS::card_shift; CardIdx_t card_index = from_card - from_hr_bot_card_index; assert(0 <= card_index && (size_t)card_index < HeapRegion::CardsPerRegion, "Must be in range."); if (G1HRRSUseSparseTable && _sparse_table.add_card(from_hrs_ind, card_index)) { if (G1RecordHRRSOops) { HeapRegionRemSet::record(hr(), from); #if HRRS_VERBOSE gclog_or_tty->print(" Added card " PTR_FORMAT " to region " "[" PTR_FORMAT "...) for ref " PTR_FORMAT ".\n", align_size_down(uintptr_t(from), CardTableModRefBS::card_size), hr()->bottom(), from); #endif } #if HRRS_VERBOSE gclog_or_tty->print_cr(" added card to sparse table."); #endif assert(contains_reference_locked(from), "We just added it!"); return; } else { #if HRRS_VERBOSE gclog_or_tty->print_cr(" [tid %d] sparse table entry " "overflow(f: %d, t: %d)", tid, from_hrs_ind, cur_hrs_ind); #endif } if (_n_fine_entries == _max_fine_entries) { prt = delete_region_table(); } else { prt = PosParPRT::alloc(from_hr); } prt->init(from_hr); PosParPRT* first_prt = _fine_grain_regions[ind]; prt->set_next(first_prt); // XXX Maybe move to init? _fine_grain_regions[ind] = prt; _n_fine_entries++; if (G1HRRSUseSparseTable) { // Transfer from sparse to fine-grain. SparsePRTEntry *sprt_entry = _sparse_table.get_entry(from_hrs_ind); assert(sprt_entry != NULL, "There should have been an entry"); for (int i = 0; i < SparsePRTEntry::cards_num(); i++) { CardIdx_t c = sprt_entry->card(i); if (c != SparsePRTEntry::NullEntry) { prt->add_card(c); } } // Now we can delete the sparse entry. bool res = _sparse_table.delete_entry(from_hrs_ind); assert(res, "It should have been there."); } } assert(prt != NULL && prt->hr() == from_hr, "consequence"); } // Note that we can't assert "prt->hr() == from_hr", because of the // possibility of concurrent reuse. But see head comment of // OtherRegionsTable for why this is OK. assert(prt != NULL, "Inv"); if (prt->should_expand(tid)) { MutexLockerEx x(&_m, Mutex::_no_safepoint_check_flag); HeapRegion* prt_hr = prt->hr(); if (prt_hr == from_hr) { // Make sure the table still corresponds to the same region prt->par_expand(); prt->add_reference(from, tid); } // else: The table has been concurrently coarsened, evicted, and // the table data structure re-used for another table. So, we // don't need to add the reference any more given that the table // has been coarsened and the whole region will be scanned anyway. } else { prt->add_reference(from, tid); } if (G1RecordHRRSOops) { HeapRegionRemSet::record(hr(), from); #if HRRS_VERBOSE gclog_or_tty->print("Added card " PTR_FORMAT " to region " "[" PTR_FORMAT "...) for ref " PTR_FORMAT ".\n", align_size_down(uintptr_t(from), CardTableModRefBS::card_size), hr()->bottom(), from); #endif } assert(contains_reference(from), "We just added it!"); } PosParPRT* OtherRegionsTable::find_region_table(size_t ind, HeapRegion* hr) const { assert(0 <= ind && ind < _max_fine_entries, "Preconditions."); PosParPRT* prt = _fine_grain_regions[ind]; while (prt != NULL && prt->hr() != hr) { prt = prt->next(); } // Loop postcondition is the method postcondition. return prt; } #define DRT_CENSUS 0 #if DRT_CENSUS static const int HistoSize = 6; static int global_histo[HistoSize] = { 0, 0, 0, 0, 0, 0 }; static int coarsenings = 0; static int occ_sum = 0; #endif jint OtherRegionsTable::_n_coarsenings = 0; PosParPRT* OtherRegionsTable::delete_region_table() { #if DRT_CENSUS int histo[HistoSize] = { 0, 0, 0, 0, 0, 0 }; const int histo_limits[] = { 1, 4, 16, 64, 256, 2048 }; #endif assert(_m.owned_by_self(), "Precondition"); assert(_n_fine_entries == _max_fine_entries, "Precondition"); PosParPRT* max = NULL; jint max_occ = 0; PosParPRT** max_prev; size_t max_ind; #if SAMPLE_FOR_EVICTION size_t i = _fine_eviction_start; for (size_t k = 0; k < _fine_eviction_sample_size; k++) { size_t ii = i; // Make sure we get a non-NULL sample. while (_fine_grain_regions[ii] == NULL) { ii++; if (ii == _max_fine_entries) ii = 0; guarantee(ii != i, "We must find one."); } PosParPRT** prev = &_fine_grain_regions[ii]; PosParPRT* cur = *prev; while (cur != NULL) { jint cur_occ = cur->occupied(); if (max == NULL || cur_occ > max_occ) { max = cur; max_prev = prev; max_ind = i; max_occ = cur_occ; } prev = cur->next_addr(); cur = cur->next(); } i = i + _fine_eviction_stride; if (i >= _n_fine_entries) i = i - _n_fine_entries; } _fine_eviction_start++; if (_fine_eviction_start >= _n_fine_entries) _fine_eviction_start -= _n_fine_entries; #else for (int i = 0; i < _max_fine_entries; i++) { PosParPRT** prev = &_fine_grain_regions[i]; PosParPRT* cur = *prev; while (cur != NULL) { jint cur_occ = cur->occupied(); #if DRT_CENSUS for (int k = 0; k < HistoSize; k++) { if (cur_occ <= histo_limits[k]) { histo[k]++; global_histo[k]++; break; } } #endif if (max == NULL || cur_occ > max_occ) { max = cur; max_prev = prev; max_ind = i; max_occ = cur_occ; } prev = cur->next_addr(); cur = cur->next(); } } #endif // XXX guarantee(max != NULL, "Since _n_fine_entries > 0"); #if DRT_CENSUS gclog_or_tty->print_cr("In a coarsening: histo of occs:"); for (int k = 0; k < HistoSize; k++) { gclog_or_tty->print_cr(" <= %4d: %5d.", histo_limits[k], histo[k]); } coarsenings++; occ_sum += max_occ; if ((coarsenings % 100) == 0) { gclog_or_tty->print_cr("\ncoarsenings = %d; global summary:", coarsenings); for (int k = 0; k < HistoSize; k++) { gclog_or_tty->print_cr(" <= %4d: %5d.", histo_limits[k], global_histo[k]); } gclog_or_tty->print_cr("Avg occ of deleted region = %6.2f.", (float)occ_sum/(float)coarsenings); } #endif // Set the corresponding coarse bit. size_t max_hrs_index = max->hr()->hrs_index(); if (!_coarse_map.at(max_hrs_index)) { _coarse_map.at_put(max_hrs_index, true); _n_coarse_entries++; #if 0 gclog_or_tty->print("Coarsened entry in region [" PTR_FORMAT "...] " "for region [" PTR_FORMAT "...] (%d coarse entries).\n", hr()->bottom(), max->hr()->bottom(), _n_coarse_entries); #endif } // Unsplice. *max_prev = max->next(); Atomic::inc(&_n_coarsenings); _n_fine_entries--; return max; } // At present, this must be called stop-world single-threaded. void OtherRegionsTable::scrub(CardTableModRefBS* ctbs, BitMap* region_bm, BitMap* card_bm) { // First eliminated garbage regions from the coarse map. if (G1RSScrubVerbose) gclog_or_tty->print_cr("Scrubbing region "SIZE_FORMAT":", hr()->hrs_index()); assert(_coarse_map.size() == region_bm->size(), "Precondition"); if (G1RSScrubVerbose) gclog_or_tty->print(" Coarse map: before = %d...", _n_coarse_entries); _coarse_map.set_intersection(*region_bm); _n_coarse_entries = _coarse_map.count_one_bits(); if (G1RSScrubVerbose) gclog_or_tty->print_cr(" after = %d.", _n_coarse_entries); // Now do the fine-grained maps. for (size_t i = 0; i < _max_fine_entries; i++) { PosParPRT* cur = _fine_grain_regions[i]; PosParPRT** prev = &_fine_grain_regions[i]; while (cur != NULL) { PosParPRT* nxt = cur->next(); // If the entire region is dead, eliminate. if (G1RSScrubVerbose) gclog_or_tty->print_cr(" For other region "SIZE_FORMAT":", cur->hr()->hrs_index()); if (!region_bm->at(cur->hr()->hrs_index())) { *prev = nxt; cur->set_next(NULL); _n_fine_entries--; if (G1RSScrubVerbose) gclog_or_tty->print_cr(" deleted via region map."); PosParPRT::free(cur); } else { // Do fine-grain elimination. if (G1RSScrubVerbose) gclog_or_tty->print(" occ: before = %4d.", cur->occupied()); cur->scrub(ctbs, card_bm); if (G1RSScrubVerbose) gclog_or_tty->print_cr(" after = %4d.", cur->occupied()); // Did that empty the table completely? if (cur->occupied() == 0) { *prev = nxt; cur->set_next(NULL); _n_fine_entries--; PosParPRT::free(cur); } else { prev = cur->next_addr(); } } cur = nxt; } } // Since we may have deleted a from_card_cache entry from the RS, clear // the FCC. clear_fcc(); } size_t OtherRegionsTable::occupied() const { // Cast away const in this case. MutexLockerEx x((Mutex*)&_m, Mutex::_no_safepoint_check_flag); size_t sum = occ_fine(); sum += occ_sparse(); sum += occ_coarse(); return sum; } size_t OtherRegionsTable::occ_fine() const { size_t sum = 0; for (size_t i = 0; i < _max_fine_entries; i++) { PosParPRT* cur = _fine_grain_regions[i]; while (cur != NULL) { sum += cur->occupied(); cur = cur->next(); } } return sum; } size_t OtherRegionsTable::occ_coarse() const { return (_n_coarse_entries * HeapRegion::CardsPerRegion); } size_t OtherRegionsTable::occ_sparse() const { return _sparse_table.occupied(); } size_t OtherRegionsTable::mem_size() const { // Cast away const in this case. MutexLockerEx x((Mutex*)&_m, Mutex::_no_safepoint_check_flag); size_t sum = 0; for (size_t i = 0; i < _max_fine_entries; i++) { PosParPRT* cur = _fine_grain_regions[i]; while (cur != NULL) { sum += cur->mem_size(); cur = cur->next(); } } sum += (sizeof(PosParPRT*) * _max_fine_entries); sum += (_coarse_map.size_in_words() * HeapWordSize); sum += (_sparse_table.mem_size()); sum += sizeof(*this) - sizeof(_sparse_table); // Avoid double counting above. return sum; } size_t OtherRegionsTable::static_mem_size() { return _from_card_cache_mem_size; } size_t OtherRegionsTable::fl_mem_size() { return PerRegionTable::fl_mem_size() + PosParPRT::fl_mem_size(); } void OtherRegionsTable::clear_fcc() { for (int i = 0; i < HeapRegionRemSet::num_par_rem_sets(); i++) { _from_card_cache[i][hr()->hrs_index()] = -1; } } void OtherRegionsTable::clear() { MutexLockerEx x(&_m, Mutex::_no_safepoint_check_flag); for (size_t i = 0; i < _max_fine_entries; i++) { PosParPRT* cur = _fine_grain_regions[i]; while (cur != NULL) { PosParPRT* nxt = cur->next(); PosParPRT::free(cur); cur = nxt; } _fine_grain_regions[i] = NULL; } _sparse_table.clear(); _coarse_map.clear(); _n_fine_entries = 0; _n_coarse_entries = 0; clear_fcc(); } void OtherRegionsTable::clear_incoming_entry(HeapRegion* from_hr) { MutexLockerEx x(&_m, Mutex::_no_safepoint_check_flag); size_t hrs_ind = from_hr->hrs_index(); size_t ind = hrs_ind & _mod_max_fine_entries_mask; if (del_single_region_table(ind, from_hr)) { assert(!_coarse_map.at(hrs_ind), "Inv"); } else { _coarse_map.par_at_put(hrs_ind, 0); } // Check to see if any of the fcc entries come from here. size_t hr_ind = hr()->hrs_index(); for (int tid = 0; tid < HeapRegionRemSet::num_par_rem_sets(); tid++) { int fcc_ent = _from_card_cache[tid][hr_ind]; if (fcc_ent != -1) { HeapWord* card_addr = (HeapWord*) (uintptr_t(fcc_ent) << CardTableModRefBS::card_shift); if (hr()->is_in_reserved(card_addr)) { // Clear the from card cache. _from_card_cache[tid][hr_ind] = -1; } } } } bool OtherRegionsTable::del_single_region_table(size_t ind, HeapRegion* hr) { assert(0 <= ind && ind < _max_fine_entries, "Preconditions."); PosParPRT** prev_addr = &_fine_grain_regions[ind]; PosParPRT* prt = *prev_addr; while (prt != NULL && prt->hr() != hr) { prev_addr = prt->next_addr(); prt = prt->next(); } if (prt != NULL) { assert(prt->hr() == hr, "Loop postcondition."); *prev_addr = prt->next(); PosParPRT::free(prt); _n_fine_entries--; return true; } else { return false; } } bool OtherRegionsTable::contains_reference(OopOrNarrowOopStar from) const { // Cast away const in this case. MutexLockerEx x((Mutex*)&_m, Mutex::_no_safepoint_check_flag); return contains_reference_locked(from); } bool OtherRegionsTable::contains_reference_locked(OopOrNarrowOopStar from) const { HeapRegion* hr = _g1h->heap_region_containing_raw(from); if (hr == NULL) return false; RegionIdx_t hr_ind = (RegionIdx_t) hr->hrs_index(); // Is this region in the coarse map? if (_coarse_map.at(hr_ind)) return true; PosParPRT* prt = find_region_table(hr_ind & _mod_max_fine_entries_mask, hr); if (prt != NULL) { return prt->contains_reference(from); } else { uintptr_t from_card = (uintptr_t(from) >> CardTableModRefBS::card_shift); uintptr_t hr_bot_card_index = uintptr_t(hr->bottom()) >> CardTableModRefBS::card_shift; assert(from_card >= hr_bot_card_index, "Inv"); CardIdx_t card_index = from_card - hr_bot_card_index; assert(0 <= card_index && (size_t)card_index < HeapRegion::CardsPerRegion, "Must be in range."); return _sparse_table.contains_card(hr_ind, card_index); } } void OtherRegionsTable::do_cleanup_work(HRRSCleanupTask* hrrs_cleanup_task) { _sparse_table.do_cleanup_work(hrrs_cleanup_task); } // Determines how many threads can add records to an rset in parallel. // This can be done by either mutator threads together with the // concurrent refinement threads or GC threads. int HeapRegionRemSet::num_par_rem_sets() { return (int)MAX2(DirtyCardQueueSet::num_par_ids() + ConcurrentG1Refine::thread_num(), ParallelGCThreads); } HeapRegionRemSet::HeapRegionRemSet(G1BlockOffsetSharedArray* bosa, HeapRegion* hr) : _bosa(bosa), _other_regions(hr) { reset_for_par_iteration(); } void HeapRegionRemSet::setup_remset_size() { // Setup sparse and fine-grain tables sizes. // table_size = base * (log(region_size / 1M) + 1) int region_size_log_mb = MAX2((int)HeapRegion::LogOfHRGrainBytes - (int)LOG_M, 0); if (FLAG_IS_DEFAULT(G1RSetSparseRegionEntries)) { G1RSetSparseRegionEntries = G1RSetSparseRegionEntriesBase * (region_size_log_mb + 1); } if (FLAG_IS_DEFAULT(G1RSetRegionEntries)) { G1RSetRegionEntries = G1RSetRegionEntriesBase * (region_size_log_mb + 1); } guarantee(G1RSetSparseRegionEntries > 0 && G1RSetRegionEntries > 0 , "Sanity"); } bool HeapRegionRemSet::claim_iter() { if (_iter_state != Unclaimed) return false; jint res = Atomic::cmpxchg(Claimed, (jint*)(&_iter_state), Unclaimed); return (res == Unclaimed); } void HeapRegionRemSet::set_iter_complete() { _iter_state = Complete; } bool HeapRegionRemSet::iter_is_complete() { return _iter_state == Complete; } void HeapRegionRemSet::init_iterator(HeapRegionRemSetIterator* iter) const { iter->initialize(this); } #ifndef PRODUCT void HeapRegionRemSet::print() const { HeapRegionRemSetIterator iter; init_iterator(&iter); size_t card_index; while (iter.has_next(card_index)) { HeapWord* card_start = G1CollectedHeap::heap()->bot_shared()->address_for_index(card_index); gclog_or_tty->print_cr(" Card " PTR_FORMAT, card_start); } // XXX if (iter.n_yielded() != occupied()) { gclog_or_tty->print_cr("Yielded disagrees with occupied:"); gclog_or_tty->print_cr(" %6d yielded (%6d coarse, %6d fine).", iter.n_yielded(), iter.n_yielded_coarse(), iter.n_yielded_fine()); gclog_or_tty->print_cr(" %6d occ (%6d coarse, %6d fine).", occupied(), occ_coarse(), occ_fine()); } guarantee(iter.n_yielded() == occupied(), "We should have yielded all the represented cards."); } #endif void HeapRegionRemSet::cleanup() { SparsePRT::cleanup_all(); } void HeapRegionRemSet::par_cleanup() { PosParPRT::par_contract_all(); } void HeapRegionRemSet::clear() { _other_regions.clear(); assert(occupied() == 0, "Should be clear."); reset_for_par_iteration(); } void HeapRegionRemSet::reset_for_par_iteration() { _iter_state = Unclaimed; _iter_claimed = 0; // It's good to check this to make sure that the two methods are in sync. assert(verify_ready_for_par_iteration(), "post-condition"); } void HeapRegionRemSet::scrub(CardTableModRefBS* ctbs, BitMap* region_bm, BitMap* card_bm) { _other_regions.scrub(ctbs, region_bm, card_bm); } //-------------------- Iteration -------------------- HeapRegionRemSetIterator:: HeapRegionRemSetIterator() : _hrrs(NULL), _g1h(G1CollectedHeap::heap()), _bosa(NULL), _sparse_iter() { } void HeapRegionRemSetIterator::initialize(const HeapRegionRemSet* hrrs) { _hrrs = hrrs; _coarse_map = &_hrrs->_other_regions._coarse_map; _fine_grain_regions = _hrrs->_other_regions._fine_grain_regions; _bosa = _hrrs->bosa(); _is = Sparse; // Set these values so that we increment to the first region. _coarse_cur_region_index = -1; _coarse_cur_region_cur_card = (HeapRegion::CardsPerRegion-1); _cur_region_cur_card = 0; _fine_array_index = -1; _fine_cur_prt = NULL; _n_yielded_coarse = 0; _n_yielded_fine = 0; _n_yielded_sparse = 0; _sparse_iter.init(&hrrs->_other_regions._sparse_table); } bool HeapRegionRemSetIterator::coarse_has_next(size_t& card_index) { if (_hrrs->_other_regions._n_coarse_entries == 0) return false; // Go to the next card. _coarse_cur_region_cur_card++; // Was the last the last card in the current region? if (_coarse_cur_region_cur_card == HeapRegion::CardsPerRegion) { // Yes: find the next region. This may leave _coarse_cur_region_index // Set to the last index, in which case there are no more coarse // regions. _coarse_cur_region_index = (int) _coarse_map->get_next_one_offset(_coarse_cur_region_index + 1); if ((size_t)_coarse_cur_region_index < _coarse_map->size()) { _coarse_cur_region_cur_card = 0; HeapWord* r_bot = _g1h->region_at(_coarse_cur_region_index)->bottom(); _cur_region_card_offset = _bosa->index_for(r_bot); } else { return false; } } // If we didn't return false above, then we can yield a card. card_index = _cur_region_card_offset + _coarse_cur_region_cur_card; return true; } void HeapRegionRemSetIterator::fine_find_next_non_null_prt() { // Otherwise, find the next bucket list in the array. _fine_array_index++; while (_fine_array_index < (int) OtherRegionsTable::_max_fine_entries) { _fine_cur_prt = _fine_grain_regions[_fine_array_index]; if (_fine_cur_prt != NULL) return; else _fine_array_index++; } assert(_fine_cur_prt == NULL, "Loop post"); } bool HeapRegionRemSetIterator::fine_has_next(size_t& card_index) { if (fine_has_next()) { _cur_region_cur_card = _fine_cur_prt->_bm.get_next_one_offset(_cur_region_cur_card + 1); } while (!fine_has_next()) { if (_cur_region_cur_card == (size_t) HeapRegion::CardsPerRegion) { _cur_region_cur_card = 0; _fine_cur_prt = _fine_cur_prt->next(); } if (_fine_cur_prt == NULL) { fine_find_next_non_null_prt(); if (_fine_cur_prt == NULL) return false; } assert(_fine_cur_prt != NULL && _cur_region_cur_card == 0, "inv."); HeapWord* r_bot = _fine_cur_prt->hr()->bottom(); _cur_region_card_offset = _bosa->index_for(r_bot); _cur_region_cur_card = _fine_cur_prt->_bm.get_next_one_offset(0); } assert(fine_has_next(), "Or else we exited the loop via the return."); card_index = _cur_region_card_offset + _cur_region_cur_card; return true; } bool HeapRegionRemSetIterator::fine_has_next() { return _fine_cur_prt != NULL && _cur_region_cur_card < HeapRegion::CardsPerRegion; } bool HeapRegionRemSetIterator::has_next(size_t& card_index) { switch (_is) { case Sparse: if (_sparse_iter.has_next(card_index)) { _n_yielded_sparse++; return true; } // Otherwise, deliberate fall-through _is = Fine; case Fine: if (fine_has_next(card_index)) { _n_yielded_fine++; return true; } // Otherwise, deliberate fall-through _is = Coarse; case Coarse: if (coarse_has_next(card_index)) { _n_yielded_coarse++; return true; } // Otherwise... break; } assert(ParallelGCThreads > 1 || n_yielded() == _hrrs->occupied(), "Should have yielded all the cards in the rem set " "(in the non-par case)."); return false; } OopOrNarrowOopStar* HeapRegionRemSet::_recorded_oops = NULL; HeapWord** HeapRegionRemSet::_recorded_cards = NULL; HeapRegion** HeapRegionRemSet::_recorded_regions = NULL; int HeapRegionRemSet::_n_recorded = 0; HeapRegionRemSet::Event* HeapRegionRemSet::_recorded_events = NULL; int* HeapRegionRemSet::_recorded_event_index = NULL; int HeapRegionRemSet::_n_recorded_events = 0; void HeapRegionRemSet::record(HeapRegion* hr, OopOrNarrowOopStar f) { if (_recorded_oops == NULL) { assert(_n_recorded == 0 && _recorded_cards == NULL && _recorded_regions == NULL, "Inv"); _recorded_oops = NEW_C_HEAP_ARRAY(OopOrNarrowOopStar, MaxRecorded); _recorded_cards = NEW_C_HEAP_ARRAY(HeapWord*, MaxRecorded); _recorded_regions = NEW_C_HEAP_ARRAY(HeapRegion*, MaxRecorded); } if (_n_recorded == MaxRecorded) { gclog_or_tty->print_cr("Filled up 'recorded' (%d).", MaxRecorded); } else { _recorded_cards[_n_recorded] = (HeapWord*)align_size_down(uintptr_t(f), CardTableModRefBS::card_size); _recorded_oops[_n_recorded] = f; _recorded_regions[_n_recorded] = hr; _n_recorded++; } } void HeapRegionRemSet::record_event(Event evnt) { if (!G1RecordHRRSEvents) return; if (_recorded_events == NULL) { assert(_n_recorded_events == 0 && _recorded_event_index == NULL, "Inv"); _recorded_events = NEW_C_HEAP_ARRAY(Event, MaxRecordedEvents); _recorded_event_index = NEW_C_HEAP_ARRAY(int, MaxRecordedEvents); } if (_n_recorded_events == MaxRecordedEvents) { gclog_or_tty->print_cr("Filled up 'recorded_events' (%d).", MaxRecordedEvents); } else { _recorded_events[_n_recorded_events] = evnt; _recorded_event_index[_n_recorded_events] = _n_recorded; _n_recorded_events++; } } void HeapRegionRemSet::print_event(outputStream* str, Event evnt) { switch (evnt) { case Event_EvacStart: str->print("Evac Start"); break; case Event_EvacEnd: str->print("Evac End"); break; case Event_RSUpdateEnd: str->print("RS Update End"); break; } } void HeapRegionRemSet::print_recorded() { int cur_evnt = 0; Event cur_evnt_kind; int cur_evnt_ind = 0; if (_n_recorded_events > 0) { cur_evnt_kind = _recorded_events[cur_evnt]; cur_evnt_ind = _recorded_event_index[cur_evnt]; } for (int i = 0; i < _n_recorded; i++) { while (cur_evnt < _n_recorded_events && i == cur_evnt_ind) { gclog_or_tty->print("Event: "); print_event(gclog_or_tty, cur_evnt_kind); gclog_or_tty->print_cr(""); cur_evnt++; if (cur_evnt < MaxRecordedEvents) { cur_evnt_kind = _recorded_events[cur_evnt]; cur_evnt_ind = _recorded_event_index[cur_evnt]; } } gclog_or_tty->print("Added card " PTR_FORMAT " to region [" PTR_FORMAT "...]" " for ref " PTR_FORMAT ".\n", _recorded_cards[i], _recorded_regions[i]->bottom(), _recorded_oops[i]); } } void HeapRegionRemSet::reset_for_cleanup_tasks() { SparsePRT::reset_for_cleanup_tasks(); } void HeapRegionRemSet::do_cleanup_work(HRRSCleanupTask* hrrs_cleanup_task) { _other_regions.do_cleanup_work(hrrs_cleanup_task); } void HeapRegionRemSet::finish_cleanup_task(HRRSCleanupTask* hrrs_cleanup_task) { SparsePRT::finish_cleanup_task(hrrs_cleanup_task); } #ifndef PRODUCT void HeapRegionRemSet::test() { os::sleep(Thread::current(), (jlong)5000, false); G1CollectedHeap* g1h = G1CollectedHeap::heap(); // Run with "-XX:G1LogRSetRegionEntries=2", so that 1 and 5 end up in same // hash bucket. HeapRegion* hr0 = g1h->region_at(0); HeapRegion* hr1 = g1h->region_at(1); HeapRegion* hr2 = g1h->region_at(5); HeapRegion* hr3 = g1h->region_at(6); HeapRegion* hr4 = g1h->region_at(7); HeapRegion* hr5 = g1h->region_at(8); HeapWord* hr1_start = hr1->bottom(); HeapWord* hr1_mid = hr1_start + HeapRegion::GrainWords/2; HeapWord* hr1_last = hr1->end() - 1; HeapWord* hr2_start = hr2->bottom(); HeapWord* hr2_mid = hr2_start + HeapRegion::GrainWords/2; HeapWord* hr2_last = hr2->end() - 1; HeapWord* hr3_start = hr3->bottom(); HeapWord* hr3_mid = hr3_start + HeapRegion::GrainWords/2; HeapWord* hr3_last = hr3->end() - 1; HeapRegionRemSet* hrrs = hr0->rem_set(); // Make three references from region 0x101... hrrs->add_reference((OopOrNarrowOopStar)hr1_start); hrrs->add_reference((OopOrNarrowOopStar)hr1_mid); hrrs->add_reference((OopOrNarrowOopStar)hr1_last); hrrs->add_reference((OopOrNarrowOopStar)hr2_start); hrrs->add_reference((OopOrNarrowOopStar)hr2_mid); hrrs->add_reference((OopOrNarrowOopStar)hr2_last); hrrs->add_reference((OopOrNarrowOopStar)hr3_start); hrrs->add_reference((OopOrNarrowOopStar)hr3_mid); hrrs->add_reference((OopOrNarrowOopStar)hr3_last); // Now cause a coarsening. hrrs->add_reference((OopOrNarrowOopStar)hr4->bottom()); hrrs->add_reference((OopOrNarrowOopStar)hr5->bottom()); // Now, does iteration yield these three? HeapRegionRemSetIterator iter; hrrs->init_iterator(&iter); size_t sum = 0; size_t card_index; while (iter.has_next(card_index)) { HeapWord* card_start = G1CollectedHeap::heap()->bot_shared()->address_for_index(card_index); gclog_or_tty->print_cr(" Card " PTR_FORMAT ".", card_start); sum++; } guarantee(sum == 11 - 3 + 2048, "Failure"); guarantee(sum == hrrs->occupied(), "Failure"); } #endif