/
githubmirror
/
jdk22
Обзор
Документация
Войти
/
githubmirror
/
jdk22
Код
Запросы
0
Пакеты
0
Релизы
0
Аналитика
Безопасность
master
src/hotspot/share/gc/parallel/psPromotionManager.cpp
351 строка
12 KB
Albert Mingkun Yang
8319205: Parallel: Reenable work stealing after JDK-8310031
02 ноя 2023, 14:26
02 ноя 2023, 14:26
792d829
Код
Авторство
О чём код?
/* * Copyright (c) 2002, 2023, 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 "classfile/javaClasses.inline.hpp" #include "gc/parallel/mutableSpace.hpp" #include "gc/parallel/parallelScavengeHeap.hpp" #include "gc/parallel/psOldGen.hpp" #include "gc/parallel/psPromotionManager.inline.hpp" #include "gc/parallel/psScavenge.inline.hpp" #include "gc/shared/continuationGCSupport.inline.hpp" #include "gc/shared/gcTrace.hpp" #include "gc/shared/preservedMarks.inline.hpp" #include "gc/shared/taskqueue.inline.hpp" #include "logging/log.hpp" #include "logging/logStream.hpp" #include "memory/allocation.inline.hpp" #include "memory/iterator.inline.hpp" #include "memory/memRegion.hpp" #include "memory/padded.inline.hpp" #include "memory/resourceArea.hpp" #include "oops/access.inline.hpp" #include "oops/compressedOops.inline.hpp" PaddedEnd<PSPromotionManager>* PSPromotionManager::_manager_array = nullptr; PSPromotionManager::PSScannerTasksQueueSet* PSPromotionManager::_stack_array_depth = nullptr; PreservedMarksSet* PSPromotionManager::_preserved_marks_set = nullptr; PSOldGen* PSPromotionManager::_old_gen = nullptr; MutableSpace* PSPromotionManager::_young_space = nullptr; void PSPromotionManager::initialize() { ParallelScavengeHeap* heap = ParallelScavengeHeap::heap(); _old_gen = heap->old_gen(); _young_space = heap->young_gen()->to_space(); const uint promotion_manager_num = ParallelGCThreads; // To prevent false sharing, we pad the PSPromotionManagers // and make sure that the first instance starts at a cache line. assert(_manager_array == nullptr, "Attempt to initialize twice"); _manager_array = PaddedArray<PSPromotionManager, mtGC>::create_unfreeable(promotion_manager_num); _stack_array_depth = new PSScannerTasksQueueSet(ParallelGCThreads); // Create and register the PSPromotionManager(s) for the worker threads. for(uint i=0; i<ParallelGCThreads; i++) { stack_array_depth()->register_queue(i, _manager_array[i].claimed_stack_depth()); } // The VMThread gets its own PSPromotionManager, which is not available // for work stealing. assert(_preserved_marks_set == nullptr, "Attempt to initialize twice"); _preserved_marks_set = new PreservedMarksSet(true /* in_c_heap */); _preserved_marks_set->init(promotion_manager_num); for (uint i = 0; i < promotion_manager_num; i += 1) { _manager_array[i].register_preserved_marks(_preserved_marks_set->get(i)); } } // Helper functions to get around the circular dependency between // psScavenge.inline.hpp and psPromotionManager.inline.hpp. bool PSPromotionManager::should_scavenge(oop* p, bool check_to_space) { return PSScavenge::should_scavenge(p, check_to_space); } bool PSPromotionManager::should_scavenge(narrowOop* p, bool check_to_space) { return PSScavenge::should_scavenge(p, check_to_space); } PSPromotionManager* PSPromotionManager::gc_thread_promotion_manager(uint index) { assert(index < ParallelGCThreads, "index out of range"); assert(_manager_array != nullptr, "Sanity"); return &_manager_array[index]; } PSPromotionManager* PSPromotionManager::vm_thread_promotion_manager() { assert(_manager_array != nullptr, "Sanity"); return &_manager_array[0]; } void PSPromotionManager::pre_scavenge() { ParallelScavengeHeap* heap = ParallelScavengeHeap::heap(); _preserved_marks_set->assert_empty(); _young_space = heap->young_gen()->to_space(); for(uint i=0; i<ParallelGCThreads; i++) { manager_array(i)->reset(); } } bool PSPromotionManager::post_scavenge(YoungGCTracer& gc_tracer) { bool promotion_failure_occurred = false; TASKQUEUE_STATS_ONLY(print_taskqueue_stats()); for (uint i = 0; i < ParallelGCThreads; i++) { PSPromotionManager* manager = manager_array(i); assert(manager->claimed_stack_depth()->is_empty(), "should be empty"); if (manager->_promotion_failed_info.has_failed()) { gc_tracer.report_promotion_failed(manager->_promotion_failed_info); promotion_failure_occurred = true; } manager->flush_labs(); manager->flush_string_dedup_requests(); } if (!promotion_failure_occurred) { // If there was no promotion failure, the preserved mark stacks // should be empty. _preserved_marks_set->assert_empty(); } return promotion_failure_occurred; } #if TASKQUEUE_STATS void PSPromotionManager::print_local_stats(outputStream* const out, uint i) const { #define FMT " " SIZE_FORMAT_W(10) out->print_cr("%3u" FMT FMT FMT FMT, i, _array_chunk_pushes, _array_chunk_steals, _arrays_chunked, _array_chunks_processed); #undef FMT } static const char* const pm_stats_hdr[] = { " ----partial array---- arrays array", "thr push steal chunked chunks", "--- ---------- ---------- ---------- ----------" }; void PSPromotionManager::print_taskqueue_stats() { if (!log_is_enabled(Trace, gc, task, stats)) { return; } Log(gc, task, stats) log; ResourceMark rm; LogStream ls(log.trace()); stack_array_depth()->print_taskqueue_stats(&ls, "Oop Queue"); const uint hlines = sizeof(pm_stats_hdr) / sizeof(pm_stats_hdr[0]); for (uint i = 0; i < hlines; ++i) ls.print_cr("%s", pm_stats_hdr[i]); for (uint i = 0; i < ParallelGCThreads; ++i) { manager_array(i)->print_local_stats(&ls, i); } } void PSPromotionManager::reset_stats() { claimed_stack_depth()->stats.reset(); _array_chunk_pushes = _array_chunk_steals = 0; _arrays_chunked = _array_chunks_processed = 0; } #endif // TASKQUEUE_STATS PSPromotionManager::PSPromotionManager() { ParallelScavengeHeap* heap = ParallelScavengeHeap::heap(); // We set the old lab's start array. _old_lab.set_start_array(old_gen()->start_array()); uint queue_size = claimed_stack_depth()->max_elems(); if (ParallelGCThreads == 1) { _target_stack_size = 0; } else { _target_stack_size = GCDrainStackTargetSize; } _array_chunk_size = ParGCArrayScanChunk; // let's choose 1.5x the chunk size _min_array_size_for_chunking = 3 * _array_chunk_size / 2; _preserved_marks = nullptr; reset(); } void PSPromotionManager::reset() { assert(stacks_empty(), "reset of non-empty stack"); // We need to get an assert in here to make sure the labs are always flushed. // Do not prefill the LAB's, save heap wastage! HeapWord* lab_base = young_space()->top(); _young_lab.initialize(MemRegion(lab_base, (size_t)0)); _young_gen_is_full = false; lab_base = old_gen()->object_space()->top(); _old_lab.initialize(MemRegion(lab_base, (size_t)0)); _old_gen_is_full = false; _promotion_failed_info.reset(); TASKQUEUE_STATS_ONLY(reset_stats()); } void PSPromotionManager::register_preserved_marks(PreservedMarks* preserved_marks) { assert(_preserved_marks == nullptr, "do not set it twice"); _preserved_marks = preserved_marks; } void PSPromotionManager::restore_preserved_marks() { _preserved_marks_set->restore(&ParallelScavengeHeap::heap()->workers()); } void PSPromotionManager::drain_stacks_depth(bool totally_drain) { const uint threshold = totally_drain ? 0 : _target_stack_size; PSScannerTasksQueue* const tq = claimed_stack_depth(); do { ScannerTask task; // Drain overflow stack first, so other threads can steal from // claimed stack while we work. while (tq->pop_overflow(task)) { if (!tq->try_push_to_taskqueue(task)) { process_popped_location_depth(task); } } while (tq->pop_local(task, threshold)) { process_popped_location_depth(task); } } while (!tq->overflow_empty()); assert(!totally_drain || tq->taskqueue_empty(), "Sanity"); assert(totally_drain || tq->size() <= _target_stack_size, "Sanity"); assert(tq->overflow_empty(), "Sanity"); } void PSPromotionManager::flush_labs() { assert(stacks_empty(), "Attempt to flush lab with live stack"); // If either promotion lab fills up, we can flush the // lab but not refill it, so check first. assert(!_young_lab.is_flushed() || _young_gen_is_full, "Sanity"); if (!_young_lab.is_flushed()) _young_lab.flush(); assert(!_old_lab.is_flushed() || _old_gen_is_full, "Sanity"); if (!_old_lab.is_flushed()) _old_lab.flush(); // Let PSScavenge know if we overflowed if (_young_gen_is_full) { PSScavenge::set_survivor_overflow(true); } } template <class T> void PSPromotionManager::process_array_chunk_work( oop obj, int start, int end) { assert(start <= end, "invariant"); T* const base = (T*)objArrayOop(obj)->base(); T* p = base + start; T* const chunk_end = base + end; while (p < chunk_end) { if (PSScavenge::should_scavenge(p)) { claim_or_forward_depth(p); } ++p; } } void PSPromotionManager::process_array_chunk(PartialArrayScanTask task) { assert(PSChunkLargeArrays, "invariant"); oop old = task.to_source_array(); assert(old->is_objArray(), "invariant"); assert(old->is_forwarded(), "invariant"); TASKQUEUE_STATS_ONLY(++_array_chunks_processed); oop const obj = old->forwardee(); int start; int const end = arrayOop(old)->length(); if (end > (int) _min_array_size_for_chunking) { // we'll chunk more start = end - _array_chunk_size; assert(start > 0, "invariant"); arrayOop(old)->set_length(start); push_depth(ScannerTask(PartialArrayScanTask(old))); TASKQUEUE_STATS_ONLY(++_array_chunk_pushes); } else { // this is the final chunk for this array start = 0; int const actual_length = arrayOop(obj)->length(); arrayOop(old)->set_length(actual_length); } if (UseCompressedOops) { process_array_chunk_work<narrowOop>(obj, start, end); } else { process_array_chunk_work<oop>(obj, start, end); } } oop PSPromotionManager::oop_promotion_failed(oop obj, markWord obj_mark) { assert(_old_gen_is_full || PromotionFailureALot, "Sanity"); // Attempt to CAS in the header. // This tests if the header is still the same as when // this started. If it is the same (i.e., no forwarding // pointer has been installed), then this thread owns // it. if (obj->forward_to_atomic(obj, obj_mark) == nullptr) { // We won any races, we "own" this object. assert(obj == obj->forwardee(), "Sanity"); _promotion_failed_info.register_copy_failure(obj->size()); ContinuationGCSupport::transform_stack_chunk(obj); push_contents(obj); // Save the markWord of promotion-failed objs in _preserved_marks for later // restoration. This way we don't have to walk the young-gen to locate // these promotion-failed objs. _preserved_marks->push_always(obj, obj_mark); } else { // We lost, someone else "owns" this object guarantee(obj->is_forwarded(), "Object must be forwarded if the cas failed."); // No unallocation to worry about. obj = obj->forwardee(); } return obj; }