/* * Copyright Amazon.com Inc. or its affiliates. All Rights Reserved. * Copyright (c) 2025, 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 "gc/shenandoah/shenandoahAgeCensus.hpp" #include "gc/shenandoah/shenandoahClosures.inline.hpp" #include "gc/shenandoah/shenandoahCollectorPolicy.hpp" #include "gc/shenandoah/shenandoahFreeSet.hpp" #include "gc/shenandoah/shenandoahGenerationalControlThread.hpp" #include "gc/shenandoah/shenandoahGenerationalEvacuationTask.hpp" #include "gc/shenandoah/shenandoahGenerationalHeap.hpp" #include "gc/shenandoah/shenandoahHeap.inline.hpp" #include "gc/shenandoah/shenandoahHeapRegion.hpp" #include "gc/shenandoah/shenandoahHeapRegionClosures.hpp" #include "gc/shenandoah/shenandoahInitLogger.hpp" #include "gc/shenandoah/shenandoahMemoryPool.hpp" #include "gc/shenandoah/shenandoahMonitoringSupport.hpp" #include "gc/shenandoah/shenandoahOldGeneration.hpp" #include "gc/shenandoah/shenandoahPhaseTimings.hpp" #include "gc/shenandoah/shenandoahRegulatorThread.hpp" #include "gc/shenandoah/shenandoahScanRemembered.inline.hpp" #include "gc/shenandoah/shenandoahUtils.hpp" #include "gc/shenandoah/shenandoahWorkerPolicy.hpp" #include "gc/shenandoah/shenandoahYoungGeneration.hpp" #include "logging/log.hpp" #include "utilities/events.hpp" class ShenandoahGenerationalInitLogger : public ShenandoahInitLogger { public: static void print() { ShenandoahGenerationalInitLogger logger; logger.print_all(); } void print_heap() override { ShenandoahInitLogger::print_heap(); ShenandoahGenerationalHeap* heap = ShenandoahGenerationalHeap::heap(); ShenandoahYoungGeneration* young = heap->young_generation(); log_info(gc, init)("Young Generation Soft Size: " EXACTFMT, EXACTFMTARGS(young->soft_max_capacity())); log_info(gc, init)("Young Generation Max: " EXACTFMT, EXACTFMTARGS(young->max_capacity())); ShenandoahOldGeneration* old = heap->old_generation(); log_info(gc, init)("Old Generation Soft Size: " EXACTFMT, EXACTFMTARGS(old->soft_max_capacity())); log_info(gc, init)("Old Generation Max: " EXACTFMT, EXACTFMTARGS(old->max_capacity())); } protected: void print_gc_specific() override { ShenandoahInitLogger::print_gc_specific(); ShenandoahGenerationalHeap* heap = ShenandoahGenerationalHeap::heap(); log_info(gc, init)("Young Heuristics: %s", heap->young_generation()->heuristics()->name()); log_info(gc, init)("Old Heuristics: %s", heap->old_generation()->heuristics()->name()); } }; size_t ShenandoahGenerationalHeap::calculate_min_plab() { return align_up(PLAB::min_size(), CardTable::card_size_in_words()); } size_t ShenandoahGenerationalHeap::calculate_max_plab() { size_t MaxTLABSizeWords = ShenandoahHeapRegion::max_tlab_size_words(); return align_down(MaxTLABSizeWords, CardTable::card_size_in_words()); } // Returns size in bytes size_t ShenandoahGenerationalHeap::unsafe_max_tlab_alloc(Thread *thread) const { return MIN2(ShenandoahHeapRegion::max_tlab_size_bytes(), young_generation()->available()); } ShenandoahGenerationalHeap::ShenandoahGenerationalHeap(ShenandoahCollectorPolicy* policy) : ShenandoahHeap(policy), _age_census(nullptr), _evac_tracker(new ShenandoahEvacuationTracker()), _min_plab_size(calculate_min_plab()), _max_plab_size(calculate_max_plab()), _regulator_thread(nullptr), _young_gen_memory_pool(nullptr), _old_gen_memory_pool(nullptr) { assert(is_aligned(_min_plab_size, CardTable::card_size_in_words()), "min_plab_size must be aligned"); assert(is_aligned(_max_plab_size, CardTable::card_size_in_words()), "max_plab_size must be aligned"); } void ShenandoahGenerationalHeap::post_initialize() { ShenandoahHeap::post_initialize(); _age_census = new ShenandoahAgeCensus(); } void ShenandoahGenerationalHeap::print_init_logger() const { ShenandoahGenerationalInitLogger logger; logger.print_all(); } void ShenandoahGenerationalHeap::print_tracing_info() const { ShenandoahHeap::print_tracing_info(); LogTarget(Info, gc, stats) lt; if (lt.is_enabled()) { LogStream ls(lt); ls.cr(); ls.cr(); evac_tracker()->print_global_on(&ls); } } void ShenandoahGenerationalHeap::initialize_heuristics() { // Initialize global generation and heuristics even in generational mode. ShenandoahHeap::initialize_heuristics(); // Max capacity is the maximum _allowed_ capacity. That is, the maximum allowed capacity // for old would be total heap - minimum capacity of young. This means the sum of the maximum // allowed for old and young could exceed the total heap size. It remains the case that the // _actual_ capacity of young + old = total. _generation_sizer.heap_size_changed(max_capacity()); size_t initial_capacity_young = _generation_sizer.max_young_size(); size_t max_capacity_young = _generation_sizer.max_young_size(); size_t initial_capacity_old = max_capacity() - max_capacity_young; size_t max_capacity_old = max_capacity() - initial_capacity_young; _young_generation = new ShenandoahYoungGeneration(max_workers(), max_capacity_young, initial_capacity_young); _old_generation = new ShenandoahOldGeneration(max_workers(), max_capacity_old, initial_capacity_old); _young_generation->initialize_heuristics(mode()); _old_generation->initialize_heuristics(mode()); } void ShenandoahGenerationalHeap::initialize_serviceability() { assert(mode()->is_generational(), "Only for the generational mode"); _young_gen_memory_pool = new ShenandoahYoungGenMemoryPool(this); _old_gen_memory_pool = new ShenandoahOldGenMemoryPool(this); cycle_memory_manager()->add_pool(_young_gen_memory_pool); cycle_memory_manager()->add_pool(_old_gen_memory_pool); stw_memory_manager()->add_pool(_young_gen_memory_pool); stw_memory_manager()->add_pool(_old_gen_memory_pool); } GrowableArray ShenandoahGenerationalHeap::memory_pools() { assert(mode()->is_generational(), "Only for the generational mode"); GrowableArray memory_pools(2); memory_pools.append(_young_gen_memory_pool); memory_pools.append(_old_gen_memory_pool); return memory_pools; } void ShenandoahGenerationalHeap::initialize_controller() { auto control_thread = new ShenandoahGenerationalControlThread(); _control_thread = control_thread; _regulator_thread = new ShenandoahRegulatorThread(control_thread); } void ShenandoahGenerationalHeap::gc_threads_do(ThreadClosure* tcl) const { if (!shenandoah_policy()->is_at_shutdown()) { ShenandoahHeap::gc_threads_do(tcl); tcl->do_thread(regulator_thread()); } } void ShenandoahGenerationalHeap::stop() { ShenandoahHeap::stop(); regulator_thread()->stop(); } bool ShenandoahGenerationalHeap::requires_barriers(stackChunkOop obj) const { if (is_idle()) { return false; } if (is_concurrent_young_mark_in_progress() && is_in_young(obj) && !marking_context()->allocated_after_mark_start(obj)) { // We are marking young, this object is in young, and it is below the TAMS return true; } if (is_in_old(obj)) { // Card marking barriers are required for objects in the old generation return true; } if (has_forwarded_objects()) { // Object may have pointers that need to be updated return true; } return false; } void ShenandoahGenerationalHeap::evacuate_collection_set(bool concurrent) { ShenandoahRegionIterator regions; ShenandoahGenerationalEvacuationTask task(this, ®ions, concurrent, false /* only promote regions */); workers()->run_task(&task); } void ShenandoahGenerationalHeap::promote_regions_in_place(bool concurrent) { ShenandoahRegionIterator regions; ShenandoahGenerationalEvacuationTask task(this, ®ions, concurrent, true /* only promote regions */); workers()->run_task(&task); } oop ShenandoahGenerationalHeap::evacuate_object(oop p, Thread* thread) { assert(thread == Thread::current(), "Expected thread parameter to be current thread."); if (ShenandoahThreadLocalData::is_oom_during_evac(thread)) { // This thread went through the OOM during evac protocol and it is safe to return // the forward pointer. It must not attempt to evacuate anymore. return ShenandoahBarrierSet::resolve_forwarded(p); } assert(ShenandoahThreadLocalData::is_evac_allowed(thread), "must be enclosed in oom-evac scope"); ShenandoahHeapRegion* r = heap_region_containing(p); assert(!r->is_humongous(), "never evacuate humongous objects"); ShenandoahAffiliation target_gen = r->affiliation(); // gc_generation() can change asynchronously and should not be used here. assert(active_generation() != nullptr, "Error"); if (active_generation()->is_young() && target_gen == YOUNG_GENERATION) { markWord mark = p->mark(); if (mark.is_marked()) { // Already forwarded. return ShenandoahBarrierSet::resolve_forwarded(p); } if (mark.has_displaced_mark_helper()) { // We don't want to deal with MT here just to ensure we read the right mark word. // Skip the potential promotion attempt for this one. } else if (r->age() + mark.age() >= age_census()->tenuring_threshold()) { oop result = try_evacuate_object(p, thread, r, OLD_GENERATION); if (result != nullptr) { return result; } // If we failed to promote this aged object, we'll fall through to code below and evacuate to young-gen. } } return try_evacuate_object(p, thread, r, target_gen); } // try_evacuate_object registers the object and dirties the associated remembered set information when evacuating // to OLD_GENERATION. oop ShenandoahGenerationalHeap::try_evacuate_object(oop p, Thread* thread, ShenandoahHeapRegion* from_region, ShenandoahAffiliation target_gen) { bool alloc_from_lab = true; bool has_plab = false; HeapWord* copy = nullptr; size_t size = ShenandoahForwarding::size(p); bool is_promotion = (target_gen == OLD_GENERATION) && from_region->is_young(); #ifdef ASSERT if (ShenandoahOOMDuringEvacALot && (os::random() & 1) == 0) { // Simulate OOM every ~2nd slow-path call copy = nullptr; } else { #endif if (UseTLAB) { switch (target_gen) { case YOUNG_GENERATION: { copy = allocate_from_gclab(thread, size); if ((copy == nullptr) && (size < ShenandoahThreadLocalData::gclab_size(thread))) { // GCLAB allocation failed because we are bumping up against the limit on young evacuation reserve. Try resetting // the desired GCLAB size and retry GCLAB allocation to avoid cascading of shared memory allocations. ShenandoahThreadLocalData::set_gclab_size(thread, PLAB::min_size()); copy = allocate_from_gclab(thread, size); // If we still get nullptr, we'll try a shared allocation below. } break; } case OLD_GENERATION: { PLAB* plab = ShenandoahThreadLocalData::plab(thread); if (plab != nullptr) { has_plab = true; copy = allocate_from_plab(thread, size, is_promotion); if ((copy == nullptr) && (size < ShenandoahThreadLocalData::plab_size(thread)) && ShenandoahThreadLocalData::plab_retries_enabled(thread)) { // PLAB allocation failed because we are bumping up against the limit on old evacuation reserve or because // the requested object does not fit within the current plab but the plab still has an "abundance" of memory, // where abundance is defined as >= ShenGenHeap::plab_min_size(). In the former case, we try shrinking the // desired PLAB size to the minimum and retry PLAB allocation to avoid cascading of shared memory allocations. if (plab->words_remaining() < plab_min_size()) { ShenandoahThreadLocalData::set_plab_size(thread, plab_min_size()); copy = allocate_from_plab(thread, size, is_promotion); // If we still get nullptr, we'll try a shared allocation below. if (copy == nullptr) { // If retry fails, don't continue to retry until we have success (probably in next GC pass) ShenandoahThreadLocalData::disable_plab_retries(thread); } } // else, copy still equals nullptr. this causes shared allocation below, preserving this plab for future needs. } } break; } default: { ShouldNotReachHere(); break; } } } if (copy == nullptr) { // If we failed to allocate in LAB, we'll try a shared allocation. if (!is_promotion || !has_plab || (size > PLAB::min_size())) { ShenandoahAllocRequest req = ShenandoahAllocRequest::for_shared_gc(size, target_gen, is_promotion); copy = allocate_memory(req); alloc_from_lab = false; } // else, we leave copy equal to nullptr, signaling a promotion failure below if appropriate. // We choose not to promote objects smaller than PLAB::min_size() by way of shared allocations, as this is too // costly. Instead, we'll simply "evacuate" to young-gen memory (using a GCLAB) and will promote in a future // evacuation pass. This condition is denoted by: is_promotion && has_plab && (size <= PLAB::min_size()) } #ifdef ASSERT } #endif if (copy == nullptr) { if (target_gen == OLD_GENERATION) { if (from_region->is_young()) { // Signal that promotion failed. Will evacuate this old object somewhere in young gen. old_generation()->handle_failed_promotion(thread, size); return nullptr; } else { // Remember that evacuation to old gen failed. We'll want to trigger a full gc to recover from this // after the evacuation threads have finished. old_generation()->handle_failed_evacuation(); } } control_thread()->handle_alloc_failure_evac(size); oom_evac_handler()->handle_out_of_memory_during_evacuation(); return ShenandoahBarrierSet::resolve_forwarded(p); } // Copy the object: NOT_PRODUCT(evac_tracker()->begin_evacuation(thread, size * HeapWordSize)); Copy::aligned_disjoint_words(cast_from_oop(p), copy, size); oop copy_val = cast_to_oop(copy); // Update the age of the evacuated object if (target_gen == YOUNG_GENERATION && is_aging_cycle()) { ShenandoahHeap::increase_object_age(copy_val, from_region->age() + 1); } // Try to install the new forwarding pointer. oop result = ShenandoahForwarding::try_update_forwardee(p, copy_val); if (result == copy_val) { // Successfully evacuated. Our copy is now the public one! // This is necessary for virtual thread support. This uses the mark word without // considering that it may now be a forwarding pointer (and could therefore crash). // Secondarily, we do not want to spend cycles relativizing stack chunks for oops // that lost the evacuation race (and will therefore not become visible). It is // safe to do this on the public copy (this is also done during concurrent mark). ContinuationGCSupport::relativize_stack_chunk(copy_val); // Record that the evacuation succeeded NOT_PRODUCT(evac_tracker()->end_evacuation(thread, size * HeapWordSize)); if (target_gen == OLD_GENERATION) { old_generation()->handle_evacuation(copy, size, from_region->is_young()); } else { // When copying to the old generation above, we don't care // about recording object age in the census stats. assert(target_gen == YOUNG_GENERATION, "Error"); // We record this census only when simulating pre-adaptive tenuring behavior, or // when we have been asked to record the census at evacuation rather than at mark if (ShenandoahGenerationalCensusAtEvac || !ShenandoahGenerationalAdaptiveTenuring) { evac_tracker()->record_age(thread, size * HeapWordSize, ShenandoahHeap::get_object_age(copy_val)); } } shenandoah_assert_correct(nullptr, copy_val); return copy_val; } else { // Failed to evacuate. We need to deal with the object that is left behind. Since this // new allocation is certainly after TAMS, it will be considered live in the next cycle. // But if it happens to contain references to evacuated regions, those references would // not get updated for this stale copy during this cycle, and we will crash while scanning // it the next cycle. if (alloc_from_lab) { // For LAB allocations, it is enough to rollback the allocation ptr. Either the next // object will overwrite this stale copy, or the filler object on LAB retirement will // do this. switch (target_gen) { case YOUNG_GENERATION: { ShenandoahThreadLocalData::gclab(thread)->undo_allocation(copy, size); break; } case OLD_GENERATION: { ShenandoahThreadLocalData::plab(thread)->undo_allocation(copy, size); if (is_promotion) { ShenandoahThreadLocalData::subtract_from_plab_promoted(thread, size * HeapWordSize); } break; } default: { ShouldNotReachHere(); break; } } } else { // For non-LAB allocations, we have no way to retract the allocation, and // have to explicitly overwrite the copy with the filler object. With that overwrite, // we have to keep the fwdptr initialized and pointing to our (stale) copy. assert(size >= ShenandoahHeap::min_fill_size(), "previously allocated object known to be larger than min_size"); fill_with_object(copy, size); shenandoah_assert_correct(nullptr, copy_val); // For non-LAB allocations, the object has already been registered } shenandoah_assert_correct(nullptr, result); return result; } } inline HeapWord* ShenandoahGenerationalHeap::allocate_from_plab(Thread* thread, size_t size, bool is_promotion) { assert(UseTLAB, "TLABs should be enabled"); PLAB* plab = ShenandoahThreadLocalData::plab(thread); HeapWord* obj; if (plab == nullptr) { assert(!thread->is_Java_thread() && !thread->is_Worker_thread(), "Performance: thread should have PLAB: %s", thread->name()); // No PLABs in this thread, fallback to shared allocation return nullptr; } else if (is_promotion && !ShenandoahThreadLocalData::allow_plab_promotions(thread)) { return nullptr; } // if plab->word_size() <= 0, thread's plab not yet initialized for this pass, so allow_plab_promotions() is not trustworthy obj = plab->allocate(size); if ((obj == nullptr) && (plab->words_remaining() < plab_min_size())) { // allocate_from_plab_slow will establish allow_plab_promotions(thread) for future invocations obj = allocate_from_plab_slow(thread, size, is_promotion); } // if plab->words_remaining() >= ShenGenHeap::heap()->plab_min_size(), just return nullptr so we can use a shared allocation if (obj == nullptr) { return nullptr; } if (is_promotion) { ShenandoahThreadLocalData::add_to_plab_promoted(thread, size * HeapWordSize); } return obj; } // Establish a new PLAB and allocate size HeapWords within it. HeapWord* ShenandoahGenerationalHeap::allocate_from_plab_slow(Thread* thread, size_t size, bool is_promotion) { // New object should fit the PLAB size assert(mode()->is_generational(), "PLABs only relevant to generational GC"); const size_t plab_min_size = this->plab_min_size(); // PLABs are aligned to card boundaries to avoid synchronization with concurrent // allocations in other PLABs. const size_t min_size = (size > plab_min_size)? align_up(size, CardTable::card_size_in_words()): plab_min_size; // Figure out size of new PLAB, using value determined at last refill. size_t cur_size = ShenandoahThreadLocalData::plab_size(thread); if (cur_size == 0) { cur_size = plab_min_size; } // Expand aggressively, doubling at each refill in this epoch, ceiling at plab_max_size() size_t future_size = MIN2(cur_size * 2, plab_max_size()); // Doubling, starting at a card-multiple, should give us a card-multiple. (Ceiling and floor // are card multiples.) assert(is_aligned(future_size, CardTable::card_size_in_words()), "Card multiple by construction, future_size: %zu" ", card_size: %zu, cur_size: %zu, max: %zu", future_size, (size_t) CardTable::card_size_in_words(), cur_size, plab_max_size()); // Record new heuristic value even if we take any shortcut. This captures // the case when moderately-sized objects always take a shortcut. At some point, // heuristics should catch up with them. Note that the requested cur_size may // not be honored, but we remember that this is the preferred size. log_debug(gc, free)("Set new PLAB size: %zu", future_size); ShenandoahThreadLocalData::set_plab_size(thread, future_size); if (cur_size < size) { // The PLAB to be allocated is still not large enough to hold the object. Fall back to shared allocation. // This avoids retiring perfectly good PLABs in order to represent a single large object allocation. log_debug(gc, free)("Current PLAB size (%zu) is too small for %zu", cur_size, size); return nullptr; } // Retire current PLAB, and allocate a new one. PLAB* plab = ShenandoahThreadLocalData::plab(thread); if (plab->words_remaining() < plab_min_size) { // Retire current PLAB. This takes care of any PLAB book-keeping. // retire_plab() registers the remnant filler object with the remembered set scanner without a lock. // Since PLABs are card-aligned, concurrent registrations in other PLABs don't interfere. retire_plab(plab, thread); size_t actual_size = 0; HeapWord* plab_buf = allocate_new_plab(min_size, cur_size, &actual_size); if (plab_buf == nullptr) { if (min_size == plab_min_size) { // Disable PLAB promotions for this thread because we cannot even allocate a minimal PLAB. This allows us // to fail faster on subsequent promotion attempts. ShenandoahThreadLocalData::disable_plab_promotions(thread); } return nullptr; } else { ShenandoahThreadLocalData::enable_plab_retries(thread); } // Since the allocated PLAB may have been down-sized for alignment, plab->allocate(size) below may still fail. if (ZeroTLAB) { // ... and clear it. Copy::zero_to_words(plab_buf, actual_size); } else { // ...and zap just allocated object. #ifdef ASSERT // Skip mangling the space corresponding to the object header to // ensure that the returned space is not considered parsable by // any concurrent GC thread. size_t hdr_size = oopDesc::header_size(); Copy::fill_to_words(plab_buf + hdr_size, actual_size - hdr_size, badHeapWordVal); #endif // ASSERT } assert(is_aligned(actual_size, CardTable::card_size_in_words()), "Align by design"); plab->set_buf(plab_buf, actual_size); if (is_promotion && !ShenandoahThreadLocalData::allow_plab_promotions(thread)) { return nullptr; } return plab->allocate(size); } else { // If there's still at least min_size() words available within the current plab, don't retire it. Let's nibble // away on this plab as long as we can. Meanwhile, return nullptr to force this particular allocation request // to be satisfied with a shared allocation. By packing more promotions into the previously allocated PLAB, we // reduce the likelihood of evacuation failures, and we reduce the need for downsizing our PLABs. return nullptr; } } HeapWord* ShenandoahGenerationalHeap::allocate_new_plab(size_t min_size, size_t word_size, size_t* actual_size) { // Align requested sizes to card-sized multiples. Align down so that we don't violate max size of TLAB. assert(is_aligned(min_size, CardTable::card_size_in_words()), "Align by design"); assert(word_size >= min_size, "Requested PLAB is too small"); ShenandoahAllocRequest req = ShenandoahAllocRequest::for_plab(min_size, word_size); // Note that allocate_memory() sets a thread-local flag to prohibit further promotions by this thread // if we are at risk of infringing on the old-gen evacuation budget. HeapWord* res = allocate_memory(req); if (res != nullptr) { *actual_size = req.actual_size(); } else { *actual_size = 0; } assert(is_aligned(res, CardTable::card_size_in_words()), "Align by design"); return res; } void ShenandoahGenerationalHeap::retire_plab(PLAB* plab, Thread* thread) { // We don't enforce limits on plab evacuations. We let it consume all available old-gen memory in order to reduce // probability of an evacuation failure. We do enforce limits on promotion, to make sure that excessive promotion // does not result in an old-gen evacuation failure. Note that a failed promotion is relatively harmless. Any // object that fails to promote in the current cycle will be eligible for promotion in a subsequent cycle. // When the plab was instantiated, its entirety was treated as if the entire buffer was going to be dedicated to // promotions. Now that we are retiring the buffer, we adjust for the reality that the plab is not entirely promotions. // 1. Some of the plab may have been dedicated to evacuations. // 2. Some of the plab may have been abandoned due to waste (at the end of the plab). size_t not_promoted = ShenandoahThreadLocalData::get_plab_actual_size(thread) - ShenandoahThreadLocalData::get_plab_promoted(thread); ShenandoahThreadLocalData::reset_plab_promoted(thread); ShenandoahThreadLocalData::set_plab_actual_size(thread, 0); if (not_promoted > 0) { old_generation()->unexpend_promoted(not_promoted); } const size_t original_waste = plab->waste(); HeapWord* const top = plab->top(); // plab->retire() overwrites unused memory between plab->top() and plab->hard_end() with a dummy object to make memory parsable. // It adds the size of this unused memory, in words, to plab->waste(). plab->retire(); if (top != nullptr && plab->waste() > original_waste && is_in_old(top)) { // If retiring the plab created a filler object, then we need to register it with our card scanner so it can // safely walk the region backing the plab. log_debug(gc)("retire_plab() is registering remnant of size %zu at " PTR_FORMAT, plab->waste() - original_waste, p2i(top)); // No lock is necessary because the PLAB memory is aligned on card boundaries. old_generation()->card_scan()->register_object_without_lock(top); } } void ShenandoahGenerationalHeap::retire_plab(PLAB* plab) { Thread* thread = Thread::current(); retire_plab(plab, thread); } ShenandoahGenerationalHeap::TransferResult ShenandoahGenerationalHeap::balance_generations() { shenandoah_assert_heaplocked_or_safepoint(); ShenandoahOldGeneration* old_gen = old_generation(); const ssize_t old_region_balance = old_gen->get_region_balance(); old_gen->set_region_balance(0); if (old_region_balance > 0) { const auto old_region_surplus = checked_cast(old_region_balance); const bool success = generation_sizer()->transfer_to_young(old_region_surplus); return TransferResult { success, old_region_surplus, "young" }; } if (old_region_balance < 0) { const auto old_region_deficit = checked_cast(-old_region_balance); const bool success = generation_sizer()->transfer_to_old(old_region_deficit); if (!success) { old_gen->handle_failed_transfer(); } return TransferResult { success, old_region_deficit, "old" }; } return TransferResult {true, 0, "none"}; } // Make sure old-generation is large enough, but no larger than is necessary, to hold mixed evacuations // and promotions, if we anticipate either. Any deficit is provided by the young generation, subject to // xfer_limit, and any surplus is transferred to the young generation. // xfer_limit is the maximum we're able to transfer from young to old. void ShenandoahGenerationalHeap::compute_old_generation_balance(size_t old_xfer_limit, size_t old_cset_regions) { // We can limit the old reserve to the size of anticipated promotions: // max_old_reserve is an upper bound on memory evacuated from old and promoted to old, // clamped by the old generation space available. // // Here's the algebra. // Let SOEP = ShenandoahOldEvacRatioPercent, // OE = old evac, // YE = young evac, and // TE = total evac = OE + YE // By definition: // SOEP/100 = OE/TE // = OE/(OE+YE) // => SOEP/(100-SOEP) = OE/((OE+YE)-OE) // componendo-dividendo: If a/b = c/d, then a/(b-a) = c/(d-c) // = OE/YE // => OE = YE*SOEP/(100-SOEP) // We have to be careful in the event that SOEP is set to 100 by the user. assert(ShenandoahOldEvacRatioPercent <= 100, "Error"); const size_t old_available = old_generation()->available(); // The free set will reserve this amount of memory to hold young evacuations const size_t young_reserve = (young_generation()->max_capacity() * ShenandoahEvacReserve) / 100; // In the case that ShenandoahOldEvacRatioPercent equals 100, max_old_reserve is limited only by xfer_limit. const double bound_on_old_reserve = old_available + old_xfer_limit + young_reserve; const double max_old_reserve = (ShenandoahOldEvacRatioPercent == 100)? bound_on_old_reserve: MIN2(double(young_reserve * ShenandoahOldEvacRatioPercent) / double(100 - ShenandoahOldEvacRatioPercent), bound_on_old_reserve); const size_t region_size_bytes = ShenandoahHeapRegion::region_size_bytes(); // Decide how much old space we should reserve for a mixed collection double reserve_for_mixed = 0; if (old_generation()->has_unprocessed_collection_candidates()) { // We want this much memory to be unfragmented in order to reliably evacuate old. This is conservative because we // may not evacuate the entirety of unprocessed candidates in a single mixed evacuation. const double max_evac_need = (double(old_generation()->unprocessed_collection_candidates_live_memory()) * ShenandoahOldEvacWaste); assert(old_available >= old_generation()->free_unaffiliated_regions() * region_size_bytes, "Unaffiliated available must be less than total available"); const double old_fragmented_available = double(old_available - old_generation()->free_unaffiliated_regions() * region_size_bytes); reserve_for_mixed = max_evac_need + old_fragmented_available; if (reserve_for_mixed > max_old_reserve) { reserve_for_mixed = max_old_reserve; } } // Decide how much space we should reserve for promotions from young size_t reserve_for_promo = 0; const size_t promo_load = old_generation()->get_promotion_potential(); const bool doing_promotions = promo_load > 0; if (doing_promotions) { // We're promoting and have a bound on the maximum amount that can be promoted assert(max_old_reserve >= reserve_for_mixed, "Sanity"); const size_t available_for_promotions = max_old_reserve - reserve_for_mixed; reserve_for_promo = MIN2((size_t)(promo_load * ShenandoahPromoEvacWaste), available_for_promotions); } // This is the total old we want to ideally reserve const size_t old_reserve = reserve_for_mixed + reserve_for_promo; assert(old_reserve <= max_old_reserve, "cannot reserve more than max for old evacuations"); // We now check if the old generation is running a surplus or a deficit. const size_t max_old_available = old_generation()->available() + old_cset_regions * region_size_bytes; if (max_old_available >= old_reserve) { // We are running a surplus, so the old region surplus can go to young const size_t old_surplus = (max_old_available - old_reserve) / region_size_bytes; const size_t unaffiliated_old_regions = old_generation()->free_unaffiliated_regions() + old_cset_regions; const size_t old_region_surplus = MIN2(old_surplus, unaffiliated_old_regions); old_generation()->set_region_balance(checked_cast(old_region_surplus)); } else { // We are running a deficit which we'd like to fill from young. // Ignore that this will directly impact young_generation()->max_capacity(), // indirectly impacting young_reserve and old_reserve. These computations are conservative. // Note that deficit is rounded up by one region. const size_t old_need = (old_reserve - max_old_available + region_size_bytes - 1) / region_size_bytes; const size_t max_old_region_xfer = old_xfer_limit / region_size_bytes; // Round down the regions we can transfer from young to old. If we're running short // on young-gen memory, we restrict the xfer. Old-gen collection activities will be // curtailed if the budget is restricted. const size_t old_region_deficit = MIN2(old_need, max_old_region_xfer); old_generation()->set_region_balance(0 - checked_cast(old_region_deficit)); } } void ShenandoahGenerationalHeap::reset_generation_reserves() { young_generation()->set_evacuation_reserve(0); old_generation()->set_evacuation_reserve(0); old_generation()->set_promoted_reserve(0); } void ShenandoahGenerationalHeap::TransferResult::print_on(const char* when, outputStream* ss) const { auto heap = ShenandoahGenerationalHeap::heap(); ShenandoahYoungGeneration* const young_gen = heap->young_generation(); ShenandoahOldGeneration* const old_gen = heap->old_generation(); const size_t young_available = young_gen->available(); const size_t old_available = old_gen->available(); ss->print_cr("After %s, %s %zu regions to %s to prepare for next gc, old available: " PROPERFMT ", young_available: " PROPERFMT, when, success? "successfully transferred": "failed to transfer", region_count, region_destination, PROPERFMTARGS(old_available), PROPERFMTARGS(young_available)); } void ShenandoahGenerationalHeap::coalesce_and_fill_old_regions(bool concurrent) { class ShenandoahGlobalCoalesceAndFill : public WorkerTask { private: ShenandoahPhaseTimings::Phase _phase; ShenandoahRegionIterator _regions; public: explicit ShenandoahGlobalCoalesceAndFill(ShenandoahPhaseTimings::Phase phase) : WorkerTask("Shenandoah Global Coalesce"), _phase(phase) {} void work(uint worker_id) override { ShenandoahWorkerTimingsTracker timer(_phase, ShenandoahPhaseTimings::ScanClusters, worker_id, true); ShenandoahHeapRegion* region; while ((region = _regions.next()) != nullptr) { // old region is not in the collection set and was not immediately trashed if (region->is_old() && region->is_active() && !region->is_humongous()) { // Reset the coalesce and fill boundary because this is a global collect // and cannot be preempted by young collects. We want to be sure the entire // region is coalesced here and does not resume from a previously interrupted // or completed coalescing. region->begin_preemptible_coalesce_and_fill(); region->oop_coalesce_and_fill(false); } } } }; ShenandoahPhaseTimings::Phase phase = concurrent ? ShenandoahPhaseTimings::conc_coalesce_and_fill : ShenandoahPhaseTimings::degen_gc_coalesce_and_fill; // This is not cancellable ShenandoahGlobalCoalesceAndFill coalesce(phase); workers()->run_task(&coalesce); old_generation()->set_parsable(true); } template class ShenandoahGenerationalUpdateHeapRefsTask : public WorkerTask { private: ShenandoahGenerationalHeap* _heap; ShenandoahRegionIterator* _regions; ShenandoahRegionChunkIterator* _work_chunks; public: explicit ShenandoahGenerationalUpdateHeapRefsTask(ShenandoahRegionIterator* regions, ShenandoahRegionChunkIterator* work_chunks) : WorkerTask("Shenandoah Update References"), _heap(ShenandoahGenerationalHeap::heap()), _regions(regions), _work_chunks(work_chunks) { bool old_bitmap_stable = _heap->old_generation()->is_mark_complete(); log_debug(gc, remset)("Update refs, scan remembered set using bitmap: %s", BOOL_TO_STR(old_bitmap_stable)); } void work(uint worker_id) { if (CONCURRENT) { ShenandoahConcurrentWorkerSession worker_session(worker_id); ShenandoahSuspendibleThreadSetJoiner stsj; do_work(worker_id); } else { ShenandoahParallelWorkerSession worker_session(worker_id); do_work(worker_id); } } private: template void do_work(uint worker_id) { T cl; if (CONCURRENT && (worker_id == 0)) { // We ask the first worker to replenish the Mutator free set by moving regions previously reserved to hold the // results of evacuation. These reserves are no longer necessary because evacuation has completed. size_t cset_regions = _heap->collection_set()->count(); // Now that evacuation is done, we can reassign any regions that had been reserved to hold the results of evacuation // to the mutator free set. At the end of GC, we will have cset_regions newly evacuated fully empty regions from // which we will be able to replenish the Collector free set and the OldCollector free set in preparation for the // next GC cycle. _heap->free_set()->move_regions_from_collector_to_mutator(cset_regions); } // If !CONCURRENT, there's no value in expanding Mutator free set ShenandoahHeapRegion* r = _regions->next(); // We update references for global, old, and young collections. ShenandoahGeneration* const gc_generation = _heap->gc_generation(); shenandoah_assert_generations_reconciled(); assert(gc_generation->is_mark_complete(), "Expected complete marking"); ShenandoahMarkingContext* const ctx = _heap->marking_context(); bool is_mixed = _heap->collection_set()->has_old_regions(); while (r != nullptr) { HeapWord* update_watermark = r->get_update_watermark(); assert(update_watermark >= r->bottom(), "sanity"); log_debug(gc)("Update refs worker " UINT32_FORMAT ", looking at region %zu", worker_id, r->index()); bool region_progress = false; if (r->is_active() && !r->is_cset()) { if (r->is_young()) { _heap->marked_object_oop_iterate(r, &cl, update_watermark); region_progress = true; } else if (r->is_old()) { if (gc_generation->is_global()) { _heap->marked_object_oop_iterate(r, &cl, update_watermark); region_progress = true; } // Otherwise, this is an old region in a young or mixed cycle. Process it during a second phase, below. // Don't bother to report pacing progress in this case. } else { // Because updating of references runs concurrently, it is possible that a FREE inactive region transitions // to a non-free active region while this loop is executing. Whenever this happens, the changing of a region's // active status may propagate at a different speed than the changing of the region's affiliation. // When we reach this control point, it is because a race has allowed a region's is_active() status to be seen // by this thread before the region's affiliation() is seen by this thread. // It's ok for this race to occur because the newly transformed region does not have any references to be // updated. assert(r->get_update_watermark() == r->bottom(), "%s Region %zu is_active but not recognized as YOUNG or OLD so must be newly transitioned from FREE", r->affiliation_name(), r->index()); } } if (region_progress && ShenandoahPacing) { _heap->pacer()->report_update_refs(pointer_delta(update_watermark, r->bottom())); } if (_heap->check_cancelled_gc_and_yield(CONCURRENT)) { return; } r = _regions->next(); } if (!gc_generation->is_global()) { // Since this is generational and not GLOBAL, we have to process the remembered set. There's no remembered // set processing if not in generational mode or if GLOBAL mode. // After this thread has exhausted its traditional update-refs work, it continues with updating refs within // remembered set. The remembered set workload is better balanced between threads, so threads that are "behind" // can catch up with other threads during this phase, allowing all threads to work more effectively in parallel. update_references_in_remembered_set(worker_id, cl, ctx, is_mixed); } } template void update_references_in_remembered_set(uint worker_id, T &cl, const ShenandoahMarkingContext* ctx, bool is_mixed) { struct ShenandoahRegionChunk assignment; ShenandoahScanRemembered* scanner = _heap->old_generation()->card_scan(); while (!_heap->check_cancelled_gc_and_yield(CONCURRENT) && _work_chunks->next(&assignment)) { // Keep grabbing next work chunk to process until finished, or asked to yield ShenandoahHeapRegion* r = assignment._r; if (r->is_active() && !r->is_cset() && r->is_old()) { HeapWord* start_of_range = r->bottom() + assignment._chunk_offset; HeapWord* end_of_range = r->get_update_watermark(); if (end_of_range > start_of_range + assignment._chunk_size) { end_of_range = start_of_range + assignment._chunk_size; } if (start_of_range >= end_of_range) { continue; } // Old region in a young cycle or mixed cycle. if (is_mixed) { if (r->is_humongous()) { // Need to examine both dirty and clean cards during mixed evac. r->oop_iterate_humongous_slice_all(&cl,start_of_range, assignment._chunk_size); } else { // Since this is mixed evacuation, old regions that are candidates for collection have not been coalesced // and filled. This will use mark bits to find objects that need to be updated. update_references_in_old_region(cl, ctx, scanner, r, start_of_range, end_of_range); } } else { // This is a young evacuation size_t cluster_size = CardTable::card_size_in_words() * ShenandoahCardCluster::CardsPerCluster; size_t clusters = assignment._chunk_size / cluster_size; assert(clusters * cluster_size == assignment._chunk_size, "Chunk assignment must align on cluster boundaries"); scanner->process_region_slice(r, assignment._chunk_offset, clusters, end_of_range, &cl, true, worker_id); } if (ShenandoahPacing) { _heap->pacer()->report_update_refs(pointer_delta(end_of_range, start_of_range)); } } } } template void update_references_in_old_region(T &cl, const ShenandoahMarkingContext* ctx, ShenandoahScanRemembered* scanner, const ShenandoahHeapRegion* r, HeapWord* start_of_range, HeapWord* end_of_range) const { // In case last object in my range spans boundary of my chunk, I may need to scan all the way to top() ShenandoahObjectToOopBoundedClosure objs(&cl, start_of_range, r->top()); // Any object that begins in a previous range is part of a different scanning assignment. Any object that // starts after end_of_range is also not my responsibility. (Either allocated during evacuation, so does // not hold pointers to from-space, or is beyond the range of my assigned work chunk.) // Find the first object that begins in my range, if there is one. Note that `p` will be set to `end_of_range` // when no live object is found in the range. HeapWord* tams = ctx->top_at_mark_start(r); HeapWord* p = get_first_object_start_word(ctx, scanner, tams, start_of_range, end_of_range); while (p < end_of_range) { // p is known to point to the beginning of marked object obj oop obj = cast_to_oop(p); objs.do_object(obj); HeapWord* prev_p = p; p += obj->size(); if (p < tams) { p = ctx->get_next_marked_addr(p, tams); // If there are no more marked objects before tams, this returns tams. Note that tams is // either >= end_of_range, or tams is the start of an object that is marked. } assert(p != prev_p, "Lack of forward progress"); } } HeapWord* get_first_object_start_word(const ShenandoahMarkingContext* ctx, ShenandoahScanRemembered* scanner, HeapWord* tams, HeapWord* start_of_range, HeapWord* end_of_range) const { HeapWord* p = start_of_range; if (p >= tams) { // We cannot use ctx->is_marked(obj) to test whether an object begins at this address. Instead, // we need to use the remembered set crossing map to advance p to the first object that starts // within the enclosing card. size_t card_index = scanner->card_index_for_addr(start_of_range); while (true) { HeapWord* first_object = scanner->first_object_in_card(card_index); if (first_object != nullptr) { p = first_object; break; } else if (scanner->addr_for_card_index(card_index + 1) < end_of_range) { card_index++; } else { // Signal that no object was found in range p = end_of_range; break; } } } else if (!ctx->is_marked(cast_to_oop(p))) { p = ctx->get_next_marked_addr(p, tams); // If there are no more marked objects before tams, this returns tams. // Note that tams is either >= end_of_range, or tams is the start of an object that is marked. } return p; } }; void ShenandoahGenerationalHeap::update_heap_references(bool concurrent) { assert(!is_full_gc_in_progress(), "Only for concurrent and degenerated GC"); const uint nworkers = workers()->active_workers(); ShenandoahRegionChunkIterator work_list(nworkers); if (concurrent) { ShenandoahGenerationalUpdateHeapRefsTask task(&_update_refs_iterator, &work_list); workers()->run_task(&task); } else { ShenandoahGenerationalUpdateHeapRefsTask task(&_update_refs_iterator, &work_list); workers()->run_task(&task); } if (ShenandoahEnableCardStats) { // Only do this if we are collecting card stats ShenandoahScanRemembered* card_scan = old_generation()->card_scan(); assert(card_scan != nullptr, "Card table must exist when card stats are enabled"); card_scan->log_card_stats(nworkers, CARD_STAT_UPDATE_REFS); } } struct ShenandoahCompositeRegionClosure { template class Closure : public ShenandoahHeapRegionClosure { private: C1 &_c1; C2 &_c2; public: Closure(C1 &c1, C2 &c2) : ShenandoahHeapRegionClosure(), _c1(c1), _c2(c2) {} void heap_region_do(ShenandoahHeapRegion* r) override { _c1.heap_region_do(r); _c2.heap_region_do(r); } bool is_thread_safe() override { return _c1.is_thread_safe() && _c2.is_thread_safe(); } }; template static Closure of(C1 &c1, C2 &c2) { return Closure(c1, c2); } }; class ShenandoahUpdateRegionAges : public ShenandoahHeapRegionClosure { private: ShenandoahMarkingContext* _ctx; public: explicit ShenandoahUpdateRegionAges(ShenandoahMarkingContext* ctx) : _ctx(ctx) { } void heap_region_do(ShenandoahHeapRegion* r) override { // Maintenance of region age must follow evacuation in order to account for // evacuation allocations within survivor regions. We consult region age during // the subsequent evacuation to determine whether certain objects need to // be promoted. if (r->is_young() && r->is_active()) { HeapWord *tams = _ctx->top_at_mark_start(r); HeapWord *top = r->top(); // Allocations move the watermark when top moves. However, compacting // objects will sometimes lower top beneath the watermark, after which, // attempts to read the watermark will assert out (watermark should not be // higher than top). if (top > tams) { // There have been allocations in this region since the start of the cycle. // Any objects new to this region must not assimilate elevated age. r->reset_age(); } else if (ShenandoahGenerationalHeap::heap()->is_aging_cycle()) { r->increment_age(); } } } bool is_thread_safe() override { return true; } }; void ShenandoahGenerationalHeap::final_update_refs_update_region_states() { ShenandoahSynchronizePinnedRegionStates pins; ShenandoahUpdateRegionAges ages(active_generation()->complete_marking_context()); auto cl = ShenandoahCompositeRegionClosure::of(pins, ages); parallel_heap_region_iterate(&cl); } void ShenandoahGenerationalHeap::complete_degenerated_cycle() { shenandoah_assert_heaplocked_or_safepoint(); if (is_concurrent_old_mark_in_progress()) { // This is still necessary for degenerated cycles because the degeneration point may occur // after final mark of the young generation. See ShenandoahConcurrentGC::op_final_update_refs for // a more detailed explanation. old_generation()->transfer_pointers_from_satb(); } // We defer generation resizing actions until after cset regions have been recycled. TransferResult result = balance_generations(); LogTarget(Info, gc, ergo) lt; if (lt.is_enabled()) { LogStream ls(lt); result.print_on("Degenerated GC", &ls); } // In case degeneration interrupted concurrent evacuation or update references, we need to clean up // transient state. Otherwise, these actions have no effect. reset_generation_reserves(); if (!old_generation()->is_parsable()) { ShenandoahGCPhase phase(ShenandoahPhaseTimings::degen_gc_coalesce_and_fill); coalesce_and_fill_old_regions(false); } } void ShenandoahGenerationalHeap::complete_concurrent_cycle() { if (!old_generation()->is_parsable()) { // Class unloading may render the card offsets unusable, so we must rebuild them before // the next remembered set scan. We _could_ let the control thread do this sometime after // the global cycle has completed and before the next young collection, but under memory // pressure the control thread may not have the time (that is, because it's running back // to back GCs). In that scenario, we would have to make the old regions parsable before // we could start a young collection. This could delay the start of the young cycle and // throw off the heuristics. entry_global_coalesce_and_fill(); } TransferResult result; { ShenandoahHeapLocker locker(lock()); result = balance_generations(); reset_generation_reserves(); } LogTarget(Info, gc, ergo) lt; if (lt.is_enabled()) { LogStream ls(lt); result.print_on("Concurrent GC", &ls); } } void ShenandoahGenerationalHeap::entry_global_coalesce_and_fill() { const char* msg = "Coalescing and filling old regions"; ShenandoahConcurrentPhase gc_phase(msg, ShenandoahPhaseTimings::conc_coalesce_and_fill); TraceCollectorStats tcs(monitoring_support()->concurrent_collection_counters()); EventMark em("%s", msg); ShenandoahWorkerScope scope(workers(), ShenandoahWorkerPolicy::calc_workers_for_conc_marking(), "concurrent coalesce and fill"); coalesce_and_fill_old_regions(true); } void ShenandoahGenerationalHeap::update_region_ages(ShenandoahMarkingContext* ctx) { ShenandoahUpdateRegionAges cl(ctx); parallel_heap_region_iterate(&cl); }