ygc post

_posts/2024-09-05-g1-gc-young-gc-post.md

@@ -1,6 +1,191 @@
 ---
 layout: post
-title:  "G1 Young GC 之清理阶段"
+title:  "G1 Young GC 之收尾阶段"
 date:   2024-09-05 11:00:00 +0200
 tags: [GC, G1]
----
+---
+
+前文 GC 复制阶段已经将所有的对象从 cset 复制到其他 region，本文收尾阶段将对一些遗留问题进行处理。收尾阶段任务很多，仅讲解一些与前文相关性比较大的任务。
+
+## 总览
+
+` allocator()->release_gc_alloc_regions` 释放对象复制时使用的 region。
+
+`post_evacuate_cleanup_1` 任务集合1。
+
+`post_evacuate_cleanup_2` 任务集合2。
+
+`_evac_failure_regions.post_collection` 重置记录失败信息的数据结构。
+
+`rebuild_free_region_list` 重新构造空闲 region 列表。
+
+`prepare_for_mutator_after_young_collection` 为 Java 线程活动准备 region，并将 Survivor 区域加入到回收集。
+
+```cpp
+void G1YoungCollector::post_evacuate_collection_set(G1EvacInfo* evacuation_info, G1ParScanThreadStateSet* per_thread_states) {
+
+  //omit process weak reference process
+
+  allocator()->release_gc_alloc_regions(evacuation_info);
+
+  post_evacuate_cleanup_1(per_thread_states);
+
+  post_evacuate_cleanup_2(per_thread_states, evacuation_info);
+
+  //concurrent gc
+  if (collector_state()->in_concurrent_start_gc()) {
+    enqueue_candidates_as_root_regions();
+  }
+
+  _evac_failure_regions.post_collection();
+
+
+  _g1h->rebuild_free_region_list();
+
+  _g1h->prepare_for_mutator_after_young_collection();
+
+  _g1h->expand_heap_after_young_collection();
+}
+```
+
+## G1ClearCardTableTask
+
+此任务是清空部分 region 的 card table。
+
+`add_all_dirty_region` 在前面文章说过代表的是 cset 所有 region。由于 cset 中的对象都被清空，那么就不存在跨 region 指针，对应的 card table 区域也需要被清空。
+
+`r->clear_cardtable()` 执行清空任务。
+
+```cpp
+new G1ClearCardTableTask(G1CollectedHeap::heap(), _all_dirty_regions, this);
+
+void do_work(uint worker_id) override {
+    const uint num_regions_per_worker = num_cards_per_worker / (uint)G1HeapRegion::CardsPerRegion;
+
+    while (_cur_dirty_regions < _regions->size()) {
+    uint next = Atomic::fetch_then_add(&_cur_dirty_regions, num_regions_per_worker);
+    uint max = MIN2(next + num_regions_per_worker, _regions->size());
+    for (uint i = next; i < max; i++) {
+        G1HeapRegion* r = _g1h->region_at(_regions->at(i));
+        r->clear_cardtable();
+    }
+    }
+}
+```
+
+## RestoreEvacFailureRegionsTask
+
+此任务处理复制失败的 region，也就说 region 上还有对象没有被转移，需要做特殊处理。
+
+`_chunk_bitmap` 用于记录被遍历的内存块。
+
+`claim_chunk` 用于同步工作线程。
+
+`zap_dead_objects` 填充活对象之间的区域。
+
+```cpp
+class G1PostEvacuateCollectionSetCleanupTask1::RestoreEvacFailureRegionsTask{
+
+  CHeapBitMap _chunk_bitmap;
+  void do_work(uint worker_id) override {
+    for (uint i = 0; i < total_chunks; i++) {
+    const uint chunk_idx = (start_chunk_idx + i) % total_chunks;
+    if (claim_chunk(chunk_idx)) {
+        process_chunk(worker_id, chunk_idx);
+        } } } 
+}
+
+ void process_chunk(uint worker_id, uint chunk_idx) {
+    //omit
+    HeapWord* first_marked_addr = bitmap->get_next_marked_addr(chunk_start, hr_top);
+    HeapWord* obj_addr = first_marked_addr;
+
+    do {
+      oop obj = cast_to_oop(obj_addr);
+      const size_t obj_size = obj->size();
+      HeapWord* const obj_end_addr = obj_addr + obj_size;
+
+      obj->init_mark();
+      hr->update_bot_for_block(obj_addr, obj_end_addr);
+
+      HeapWord* next_marked_obj_addr = bitmap->get_next_marked_addr(obj_end_addr, hr_top);
+      garbage_words += zap_dead_objects(hr, obj_end_addr, next_marked_obj_addr);
+      obj_addr = next_marked_obj_addr;
+    } while (obj_addr < chunk_end);
+
+ }
+```
+
+## RestorePreservedMarksTask
+
+对于转移失败的对象，并且 markword 被存储的需要重新恢复。
+
+```cpp
+
+class G1PostEvacuateCollectionSetCleanupTask2::RestorePreservedMarksTask : public G1AbstractSubTask {
+  WorkerTask* _task;
+public:
+  RestorePreservedMarksTask(PreservedMarksSet* preserved_marks) :
+    _task(preserved_marks->create_task()) { }
+
+  void do_work(uint worker_id) override { _task->work(worker_id); }
+};
+
+WorkerTask* PreservedMarksSet::create_task() {
+  return new RestorePreservedMarksTask(this);
+}
+//_preserved_marks_set->get(task_id)->restore_and_increment(&_total_size);
+//restore()
+void PreservedMarks::restore() {
+  while (!_stack.is_empty()) {
+    const PreservedMark elem = _stack.pop();
+    elem.set_mark(); //_o->set_mark(_m);
+  }
+  assert_empty();
+}
+```
+
+## ProcessEvacuationFailedRegionsTask
+
+在 Young GC 阶段清空 markbit。
+
+## RedirtyLoggedCardsTask
+
+处理 dirty card 任务对象，除了在 cset 中，并且没有对象存在的 reigon 都会被处理
+
+```cpp
+void do_card_ptr(CardValue* card_ptr) override {
+G1HeapRegion* hr = region_for_card(card_ptr);
+    // Should only dirty cards in regions that won't be freed.
+    if (!will_become_free(hr)) {
+      *card_ptr = G1CardTable::dirty_card_val();
+      _num_dirtied++;
+    }
+}
+
+bool will_become_free(G1HeapRegion* hr) const {
+  // A region will be freed by during the FreeCollectionSet phase if the region is in the
+  // collection set and has not had an evacuation failure.
+  return _g1h->is_in_cset(hr) && !_evac_failure_regions->contains(hr->hrm_index());
+}
+```
+
+## FreeCollectionSetTask
+
+释放 cset，对于还有对象没有处理的 region ，如果存活对象小于某个阈值则需要加入到 cset 候选中，反之不处理。
+
+对于清空的 reigon 则加入 `free_list` 中。
+
+```cpp
+virtual bool do_heap_region(G1HeapRegion* r) {
+    if (_evac_failure_regions->contains(r->hrm_index())) {
+      handle_failed_region(r);
+    } else {
+      handle_evacuated_region(r);
+    }
+}
+```
+
+## 总结
+
+本文简单的介绍了 G1 的收尾阶段，对其中感兴趣的内容读者可自行深入阅读源码。

_posts/2024-09-10-g1-gc-young-gc-iter-obj.md

@@ -0,0 +1,6 @@
+---
+layout: post
+title:  "JVM 对象遍历"
+date:   2024-09-05 11:00:00 +0200
+tags: [GC, G1]
+---