8138920: Refactor the sampling thread from ConcurrentG1RefineThread

Helps enable running without concurrent refinement threads

Reviewed-by: brutisso, pliden

hotspot/src/share/vm/gc/g1/concurrentG1Refine.cpp

@@ -30,7 +30,8 @@
 #include "runtime/java.hpp"
 
 ConcurrentG1Refine::ConcurrentG1Refine(G1CollectedHeap* g1h) :
-  _threads(NULL), _n_threads(0),
+  _threads(NULL),
+  _sample_thread(NULL),
   _hot_card_cache(g1h)
 {
   // Ergonomically select initial concurrent refinement parameters
@@ -58,12 +59,10 @@ ConcurrentG1Refine* ConcurrentG1Refine::create(G1CollectedHeap* g1h, CardTableEn
     return NULL;
   }
   cg1r->_n_worker_threads = thread_num();
-  // We need one extra thread to do the young gen rset size sampling.
-  cg1r->_n_threads = cg1r->_n_worker_threads + 1;
 
   cg1r->reset_threshold_step();
 
-  cg1r->_threads = NEW_C_HEAP_ARRAY_RETURN_NULL(ConcurrentG1RefineThread*, cg1r->_n_threads, mtGC);
+  cg1r->_threads = NEW_C_HEAP_ARRAY_RETURN_NULL(ConcurrentG1RefineThread*, cg1r->_n_worker_threads, mtGC);
   if (cg1r->_threads == NULL) {
     *ecode = JNI_ENOMEM;
     vm_shutdown_during_initialization("Could not allocate an array for ConcurrentG1RefineThread");
@@ -73,7 +72,7 @@ ConcurrentG1Refine* ConcurrentG1Refine::create(G1CollectedHeap* g1h, CardTableEn
   uint worker_id_offset = DirtyCardQueueSet::num_par_ids();
 
   ConcurrentG1RefineThread *next = NULL;
-  for (uint i = cg1r->_n_threads - 1; i != UINT_MAX; i--) {
+  for (uint i = cg1r->_n_worker_threads - 1; i != UINT_MAX; i--) {
     ConcurrentG1RefineThread* t = new ConcurrentG1RefineThread(cg1r, next, refine_closure, worker_id_offset, i);
     assert(t != NULL, "Conc refine should have been created");
     if (t->osthread() == NULL) {
@@ -86,6 +85,14 @@ ConcurrentG1Refine* ConcurrentG1Refine::create(G1CollectedHeap* g1h, CardTableEn
     cg1r->_threads[i] = t;
     next = t;
   }
+
+  cg1r->_sample_thread = new G1YoungRemSetSamplingThread();
+  if (cg1r->_sample_thread->osthread() == NULL) {
+    *ecode = JNI_ENOMEM;
+    vm_shutdown_during_initialization("Could not create G1YoungRemSetSamplingThread");
+    return NULL;
+  }
+
   *ecode = JNI_OK;
   return cg1r;
 }
@@ -103,45 +110,37 @@ void ConcurrentG1Refine::init(G1RegionToSpaceMapper* card_counts_storage) {
 }
 
 void ConcurrentG1Refine::stop() {
-  if (_threads != NULL) {
+  for (uint i = 0; i < _n_worker_threads; i++) {
-    for (uint i = 0; i < _n_threads; i++) {
     _threads[i]->stop();
   }
-  }
+  _sample_thread->stop();
 }
 
 void ConcurrentG1Refine::reinitialize_threads() {
   reset_threshold_step();
-  if (_threads != NULL) {
+  for (uint i = 0; i < _n_worker_threads; i++) {
-    for (uint i = 0; i < _n_threads; i++) {
     _threads[i]->initialize();
   }
-  }
 }
 
 ConcurrentG1Refine::~ConcurrentG1Refine() {
-  if (_threads != NULL) {
+  for (uint i = 0; i < _n_worker_threads; i++) {
-    for (uint i = 0; i < _n_threads; i++) {
     delete _threads[i];
   }
   FREE_C_HEAP_ARRAY(ConcurrentG1RefineThread*, _threads);
-  }
+
+  delete _sample_thread;
 }
 
 void ConcurrentG1Refine::threads_do(ThreadClosure *tc) {
-  if (_threads != NULL) {
+  worker_threads_do(tc);
-    for (uint i = 0; i < _n_threads; i++) {
+  tc->do_thread(_sample_thread);
-      tc->do_thread(_threads[i]);
-    }
-  }
 }
 
 void ConcurrentG1Refine::worker_threads_do(ThreadClosure * tc) {
-  if (_threads != NULL) {
   for (uint i = 0; i < worker_thread_num(); i++) {
     tc->do_thread(_threads[i]);
   }
-  }
 }
 
 uint ConcurrentG1Refine::thread_num() {
@@ -149,12 +148,10 @@ uint ConcurrentG1Refine::thread_num() {
 }
 
 void ConcurrentG1Refine::print_worker_threads_on(outputStream* st) const {
-  for (uint i = 0; i < _n_threads; ++i) {
+  for (uint i = 0; i < _n_worker_threads; ++i) {
     _threads[i]->print_on(st);
     st->cr();
   }
-}
+  _sample_thread->print_on(st);
-
+  st->cr();
-ConcurrentG1RefineThread * ConcurrentG1Refine::sampling_thread() const {
-  return _threads[worker_thread_num()];
 }

hotspot/src/share/vm/gc/g1/concurrentG1Refine.hpp

@@ -26,6 +26,7 @@
 #define SHARE_VM_GC_G1_CONCURRENTG1REFINE_HPP
 
 #include "gc/g1/g1HotCardCache.hpp"
+#include "gc/g1/g1YoungRemSetSamplingThread.hpp"
 #include "memory/allocation.hpp"
 #include "runtime/thread.hpp"
 #include "utilities/globalDefinitions.hpp"
@@ -39,8 +40,9 @@ class G1RemSet;
 class DirtyCardQueue;
 
 class ConcurrentG1Refine: public CHeapObj<mtGC> {
+  G1YoungRemSetSamplingThread* _sample_thread;
+
   ConcurrentG1RefineThread** _threads;
-  uint _n_threads;
   uint _n_worker_threads;
  /*
   * The value of the update buffer queue length falls into one of 3 zones:
@@ -91,8 +93,8 @@ class ConcurrentG1Refine: public CHeapObj<mtGC> {
   // Iterate over all worker refinement threads
   void worker_threads_do(ThreadClosure * tc);
 
-  // The RS sampling thread
+  // The RS sampling thread has nothing to do with refinement, but is here for now.
-  ConcurrentG1RefineThread * sampling_thread() const;
+  G1YoungRemSetSamplingThread * sampling_thread() const { return _sample_thread; }
 
   static uint thread_num();
 
@@ -106,7 +108,6 @@ class ConcurrentG1Refine: public CHeapObj<mtGC> {
   int yellow_zone() const     { return _yellow_zone; }
   int red_zone() const        { return _red_zone;    }
 
-  uint total_thread_num() const  { return _n_threads;        }
   uint worker_thread_num() const { return _n_worker_threads; }
 
   int thread_threshold_step() const { return _thread_threshold_step; }

hotspot/src/share/vm/gc/g1/concurrentG1RefineThread.cpp

@@ -50,9 +50,8 @@ ConcurrentG1RefineThread(ConcurrentG1Refine* cg1r, ConcurrentG1RefineThread *nex
   // Each thread has its own monitor. The i-th thread is responsible for signaling
   // to thread i+1 if the number of buffers in the queue exceeds a threshold for this
   // thread. Monitors are also used to wake up the threads during termination.
-  // The 0th worker in notified by mutator threads and has a special monitor.
+  // The 0th (primary) worker is notified by mutator threads and has a special monitor.
-  // The last worker is used for young gen rset size sampling.
+  if (!is_primary()) {
-  if (worker_id > 0) {
     _monitor = new Monitor(Mutex::nonleaf, "Refinement monitor", true,
                            Monitor::_safepoint_check_never);
   } else {
@@ -66,61 +65,11 @@ ConcurrentG1RefineThread(ConcurrentG1Refine* cg1r, ConcurrentG1RefineThread *nex
 }
 
 void ConcurrentG1RefineThread::initialize() {
-  if (_worker_id < cg1r()->worker_thread_num()) {
   // Current thread activation threshold
   _threshold = MIN2<int>(cg1r()->thread_threshold_step() * (_worker_id + 1) + cg1r()->green_zone(),
                          cg1r()->yellow_zone());
   // A thread deactivates once the number of buffer reached a deactivation threshold
   _deactivation_threshold = MAX2<int>(_threshold - cg1r()->thread_threshold_step(), cg1r()->green_zone());
-  } else {
-    set_active(true);
-  }
-}
-
-void ConcurrentG1RefineThread::sample_young_list_rs_lengths() {
-  SuspendibleThreadSetJoiner sts_join;
-  G1CollectedHeap* g1h = G1CollectedHeap::heap();
-  G1CollectorPolicy* g1p = g1h->g1_policy();
-  if (g1p->adaptive_young_list_length()) {
-    int regions_visited = 0;
-    g1h->young_list()->rs_length_sampling_init();
-    while (g1h->young_list()->rs_length_sampling_more()) {
-      g1h->young_list()->rs_length_sampling_next();
-      ++regions_visited;
-
-      // we try to yield every time we visit 10 regions
-      if (regions_visited == 10) {
-        if (sts_join.should_yield()) {
-          sts_join.yield();
-          // we just abandon the iteration
-          break;
-        }
-        regions_visited = 0;
-      }
-    }
-
-    g1p->revise_young_list_target_length_if_necessary();
-  }
-}
-
-void ConcurrentG1RefineThread::run_young_rs_sampling() {
-  DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
-  _vtime_start = os::elapsedVTime();
-  while(!_should_terminate) {
-    sample_young_list_rs_lengths();
-
-    if (os::supports_vtime()) {
-      _vtime_accum = (os::elapsedVTime() - _vtime_start);
-    } else {
-      _vtime_accum = 0.0;
-    }
-
-    MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
-    if (_should_terminate) {
-      break;
-    }
-    _monitor->wait(Mutex::_no_safepoint_check_flag, G1ConcRefinementServiceIntervalMillis);
-  }
 }
 
 void ConcurrentG1RefineThread::wait_for_completed_buffers() {
@@ -133,12 +82,12 @@ void ConcurrentG1RefineThread::wait_for_completed_buffers() {
 
 bool ConcurrentG1RefineThread::is_active() {
   DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
-  return _worker_id > 0 ? _active : dcqs.process_completed_buffers();
+  return is_primary() ? dcqs.process_completed_buffers() : _active;
 }
 
 void ConcurrentG1RefineThread::activate() {
   MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
-  if (_worker_id > 0) {
+  if (!is_primary()) {
     if (G1TraceConcRefinement) {
       DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
       gclog_or_tty->print_cr("G1-Refine-activated worker %d, on threshold %d, current %d",
@@ -154,7 +103,7 @@ void ConcurrentG1RefineThread::activate() {
 
 void ConcurrentG1RefineThread::deactivate() {
   MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
-  if (_worker_id > 0) {
+  if (!is_primary()) {
     if (G1TraceConcRefinement) {
       DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
       gclog_or_tty->print_cr("G1-Refine-deactivated worker %d, off threshold %d, current %d",
@@ -171,25 +120,24 @@ void ConcurrentG1RefineThread::run() {
   initialize_in_thread();
   wait_for_universe_init();
 
-  if (_worker_id >= cg1r()->worker_thread_num()) {
+  run_service();
-    run_young_rs_sampling();
+
   terminate();
-    return;
+}
-  }
 
+void ConcurrentG1RefineThread::run_service() {
   _vtime_start = os::elapsedVTime();
-  while (!_should_terminate) {
-    DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
 
+  while (!_should_terminate) {
     // Wait for work
     wait_for_completed_buffers();
-
     if (_should_terminate) {
       break;
     }
 
     {
       SuspendibleThreadSetJoiner sts_join;
+      DirtyCardQueueSet& dcqs = JavaThread::dirty_card_queue_set();
 
       do {
         int curr_buffer_num = (int)dcqs.completed_buffers_num();
@@ -199,7 +147,7 @@ void ConcurrentG1RefineThread::run() {
           dcqs.set_completed_queue_padding(0);
         }
 
-        if (_worker_id > 0 && curr_buffer_num <= _deactivation_threshold) {
+        if (!is_primary() && curr_buffer_num <= _deactivation_threshold) {
           // If the number of the buffer has fallen below our threshold
           // we should deactivate. The predecessor will reactivate this
           // thread should the number of the buffers cross the threshold again.
@@ -225,8 +173,10 @@ void ConcurrentG1RefineThread::run() {
       _vtime_accum = 0.0;
     }
   }
-  assert(_should_terminate, "just checking");
+
-  terminate();
+  if (G1TraceConcRefinement) {
+    gclog_or_tty->print_cr("G1-Refine-stop");
+  }
 }
 
 void ConcurrentG1RefineThread::stop() {
@@ -236,10 +186,7 @@ void ConcurrentG1RefineThread::stop() {
     _should_terminate = true;
   }
 
-  {
+  stop_service();
-    MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
-    _monitor->notify();
-  }
 
   {
     MutexLockerEx mu(Terminator_lock);
@@ -247,8 +194,9 @@ void ConcurrentG1RefineThread::stop() {
       Terminator_lock->wait();
     }
   }
-  if (G1TraceConcRefinement) {
-    gclog_or_tty->print_cr("G1-Refine-stop");
-  }
 }
 
+void ConcurrentG1RefineThread::stop_service() {
+  MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
+  _monitor->notify();
+}

hotspot/src/share/vm/gc/g1/concurrentG1RefineThread.hpp

@@ -31,14 +31,14 @@
 class CardTableEntryClosure;
 class ConcurrentG1Refine;
 
-// The G1 Concurrent Refinement Thread (could be several in the future).
+// One or more G1 Concurrent Refinement Threads may be active if concurrent
-
+// refinement is in progress.
 class ConcurrentG1RefineThread: public ConcurrentGCThread {
   friend class VMStructs;
   friend class G1CollectedHeap;
 
   double _vtime_start;  // Initial virtual time.
-  double _vtime_accum;  // Initial virtual time.
+  double _vtime_accum;  // Accumulated virtual time.
   uint _worker_id;
   uint _worker_id_offset;
 
@@ -59,8 +59,6 @@ class ConcurrentG1RefineThread: public ConcurrentGCThread {
   // This thread deactivation threshold
   int _deactivation_threshold;
 
-  void sample_young_list_rs_lengths();
-  void run_young_rs_sampling();
   void wait_for_completed_buffers();
 
   void set_active(bool x) { _active = x; }
@@ -68,6 +66,11 @@ class ConcurrentG1RefineThread: public ConcurrentGCThread {
   void activate();
   void deactivate();
 
+  bool is_primary() { return (_worker_id == 0); }
+
+  void run_service();
+  void stop_service();
+
 public:
   virtual void run();
   // Constructor

hotspot/src/share/vm/gc/g1/concurrentMarkThread.cpp

@@ -92,15 +92,31 @@ void ConcurrentMarkThread::cm_log(bool doit, bool join_sts, const char* fmt, ...
   }
 }
 
+// Marking pauses can be scheduled flexibly, so we might delay marking to meet MMU.
+void ConcurrentMarkThread::delay_to_keep_mmu(G1CollectorPolicy* g1_policy, bool remark) {
+  if (g1_policy->adaptive_young_list_length()) {
+    double now = os::elapsedTime();
+    double prediction_ms = remark ? g1_policy->predict_remark_time_ms()
+                                  : g1_policy->predict_cleanup_time_ms();
+    G1MMUTracker *mmu_tracker = g1_policy->mmu_tracker();
+    jlong sleep_time_ms = mmu_tracker->when_ms(now, prediction_ms);
+    os::sleep(this, sleep_time_ms, false);
+  }
+}
 void ConcurrentMarkThread::run() {
   initialize_in_thread();
-  _vtime_start = os::elapsedVTime();
   wait_for_universe_init();
 
+  run_service();
+
+  terminate();
+}
+
+void ConcurrentMarkThread::run_service() {
+  _vtime_start = os::elapsedVTime();
+
   G1CollectedHeap* g1h = G1CollectedHeap::heap();
   G1CollectorPolicy* g1_policy = g1h->g1_policy();
-  G1MMUTracker *mmu_tracker = g1_policy->mmu_tracker();
-  Thread *current_thread = Thread::current();
 
   while (!_should_terminate) {
     // wait until started is set.
@@ -141,12 +157,7 @@ void ConcurrentMarkThread::run() {
         double mark_end_sec = os::elapsedTime();
         _vtime_mark_accum += (mark_end_time - cycle_start);
         if (!cm()->has_aborted()) {
-          if (g1_policy->adaptive_young_list_length()) {
+          delay_to_keep_mmu(g1_policy, true /* remark */);
-            double now = os::elapsedTime();
-            double remark_prediction_ms = g1_policy->predict_remark_time_ms();
-            jlong sleep_time_ms = mmu_tracker->when_ms(now, remark_prediction_ms);
-            os::sleep(current_thread, sleep_time_ms, false);
-          }
 
           cm_log(G1Log::fine(), true, "[GC concurrent-mark-end, %1.7lf secs]", mark_end_sec - mark_start_sec);
 
@@ -167,12 +178,7 @@ void ConcurrentMarkThread::run() {
       _vtime_accum = (end_time - _vtime_start);
 
       if (!cm()->has_aborted()) {
-        if (g1_policy->adaptive_young_list_length()) {
+        delay_to_keep_mmu(g1_policy, false /* cleanup */);
-          double now = os::elapsedTime();
-          double cleanup_prediction_ms = g1_policy->predict_cleanup_time_ms();
-          jlong sleep_time_ms = mmu_tracker->when_ms(now, cleanup_prediction_ms);
-          os::sleep(current_thread, sleep_time_ms, false);
-        }
 
         CMCleanUp cl_cl(_cm);
         VM_CGC_Operation op(&cl_cl, "GC cleanup", false /* needs_pll */);
@@ -272,9 +278,6 @@ void ConcurrentMarkThread::run() {
       g1h->register_concurrent_cycle_end();
     }
   }
-  assert(_should_terminate, "just checking");
-
-  terminate();
 }
 
 void ConcurrentMarkThread::stop() {
@@ -283,10 +286,7 @@ void ConcurrentMarkThread::stop() {
     _should_terminate = true;
   }
 
-  {
+  stop_service();
-    MutexLockerEx ml(CGC_lock, Mutex::_no_safepoint_check_flag);
-    CGC_lock->notify_all();
-  }
 
   {
     MutexLockerEx ml(Terminator_lock);
@@ -296,6 +296,11 @@ void ConcurrentMarkThread::stop() {
   }
 }
 
+void ConcurrentMarkThread::stop_service() {
+  MutexLockerEx ml(CGC_lock, Mutex::_no_safepoint_check_flag);
+  CGC_lock->notify_all();
+}
+
 void ConcurrentMarkThread::sleepBeforeNextCycle() {
   // We join here because we don't want to do the "shouldConcurrentMark()"
   // below while the world is otherwise stopped.

hotspot/src/share/vm/gc/g1/concurrentMarkThread.hpp

@@ -27,11 +27,11 @@
 
 #include "gc/shared/concurrentGCThread.hpp"
 
-// The Concurrent Mark GC Thread (could be several in the future).
+// The Concurrent Mark GC Thread triggers the parallel CMConcurrentMarkingTasks
-// This is copied from the Concurrent Mark Sweep GC Thread
+// as well as handling various marking cleanup.
-// Still under construction.
 
 class ConcurrentMark;
+class G1CollectorPolicy;
 
 class ConcurrentMarkThread: public ConcurrentGCThread {
   friend class VMStructs;
@@ -57,6 +57,10 @@ class ConcurrentMarkThread: public ConcurrentGCThread {
   volatile State _state;
 
   void sleepBeforeNextCycle();
+  void delay_to_keep_mmu(G1CollectorPolicy* g1_policy, bool remark);
+
+  void run_service();
+  void stop_service();
 
   static SurrogateLockerThread*         _slt;
 
@@ -67,9 +71,9 @@ class ConcurrentMarkThread: public ConcurrentGCThread {
   static void makeSurrogateLockerThread(TRAPS);
   static SurrogateLockerThread* slt() { return _slt; }
 
-  // Total virtual time so far.
+  // Total virtual time so far for this thread and concurrent marking tasks.
   double vtime_accum();
-  // Marking virtual time so far
+  // Marking virtual time so far this thread and concurrent marking tasks.
   double vtime_mark_accum();
 
   ConcurrentMark* cm()     { return _cm; }

hotspot/src/share/vm/gc/g1/g1YoungRemSetSamplingThread.cpp

@@ -0,0 +1,122 @@
+/*
+ * Copyright (c) 2015, 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/g1/g1CollectedHeap.inline.hpp"
+#include "gc/g1/g1CollectorPolicy.hpp"
+#include "gc/g1/g1YoungRemSetSamplingThread.hpp"
+#include "gc/g1/suspendibleThreadSet.hpp"
+#include "runtime/mutexLocker.hpp"
+
+void G1YoungRemSetSamplingThread::run() {
+  initialize_in_thread();
+  wait_for_universe_init();
+
+  run_service();
+
+  terminate();
+}
+
+void G1YoungRemSetSamplingThread::stop() {
+  // it is ok to take late safepoints here, if needed
+  {
+    MutexLockerEx mu(Terminator_lock);
+    _should_terminate = true;
+  }
+
+  stop_service();
+
+  {
+    MutexLockerEx mu(Terminator_lock);
+    while (!_has_terminated) {
+      Terminator_lock->wait();
+    }
+  }
+}
+
+G1YoungRemSetSamplingThread::G1YoungRemSetSamplingThread() : ConcurrentGCThread() {
+  _monitor = new Monitor(Mutex::nonleaf,
+                         "G1YoungRemSetSamplingThread monitor",
+                         true,
+                         Monitor::_safepoint_check_never);
+
+  create_and_start();
+
+  set_name("G1 Young RemSet Sampling");
+}
+
+void G1YoungRemSetSamplingThread::sleep_before_next_cycle() {
+  MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
+  if (!_should_terminate) {
+    intx waitms = G1ConcRefinementServiceIntervalMillis; // 300, really should be?
+    _monitor->wait(Mutex::_no_safepoint_check_flag, waitms);
+  }
+}
+
+void G1YoungRemSetSamplingThread::run_service() {
+  double vtime_start = os::elapsedVTime();
+
+  while (!_should_terminate) {
+    sample_young_list_rs_lengths();
+
+    if (os::supports_vtime()) {
+      _vtime_accum = (os::elapsedVTime() - vtime_start);
+    } else {
+      _vtime_accum = 0.0;
+    }
+
+    sleep_before_next_cycle();
+  }
+}
+
+void G1YoungRemSetSamplingThread::stop_service() {
+  MutexLockerEx x(_monitor, Mutex::_no_safepoint_check_flag);
+  _monitor->notify();
+}
+
+void G1YoungRemSetSamplingThread::sample_young_list_rs_lengths() {
+  SuspendibleThreadSetJoiner sts;
+  G1CollectedHeap* g1h = G1CollectedHeap::heap();
+  G1CollectorPolicy* g1p = g1h->g1_policy();
+  if (g1p->adaptive_young_list_length()) {
+    int regions_visited = 0;
+    g1h->young_list()->rs_length_sampling_init();
+    while (g1h->young_list()->rs_length_sampling_more()) {
+      g1h->young_list()->rs_length_sampling_next();
+      ++regions_visited;
+
+      // we try to yield every time we visit 10 regions
+      if (regions_visited == 10) {
+        if (sts.should_yield()) {
+          sts.yield();
+          // we just abandon the iteration
+          break;
+        }
+        regions_visited = 0;
+      }
+    }
+
+    g1p->revise_young_list_target_length_if_necessary();
+  }
+}

hotspot/src/share/vm/gc/g1/g1YoungRemSetSamplingThread.hpp

@@ -0,0 +1,63 @@
+/*
+ * Copyright (c) 2015, 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.
+ *
+ */
+
+#ifndef SHARE_VM_GC_G1_G1YOUNGREMSETSAMPLINGTHREAD_HPP
+#define SHARE_VM_GC_G1_G1YOUNGREMSETSAMPLINGTHREAD_HPP
+
+#include "gc/shared/concurrentGCThread.hpp"
+
+// The G1YoungRemSetSamplingThread is used to re-assess the validity of
+// the prediction for the remembered set lengths of the young generation.
+//
+// At the end of the GC G1 determines the length of the young gen based on
+// how much time the next GC can take, and when the next GC may occur
+// according to the MMU.
+//
+// The assumption is that a significant part of the GC is spent on scanning
+// the remembered sets (and many other components), so this thread constantly
+// reevaluates the prediction for the remembered set scanning costs, and potentially
+// G1CollectorPolicy resizes the young gen. This may do a premature GC or even
+// increase the young gen size to keep pause time length goal.
+class G1YoungRemSetSamplingThread: public ConcurrentGCThread {
+private:
+  Monitor* _monitor;
+
+  void sample_young_list_rs_lengths();
+
+  void run_service();
+  void stop_service();
+
+  void sleep_before_next_cycle();
+
+  double _vtime_accum;  // Accumulated virtual time.
+
+public:
+  G1YoungRemSetSamplingThread();
+  double vtime_accum() { return _vtime_accum; }
+
+  virtual void run();
+  void stop();
+};
+
+#endif /* SHARE_VM_GC_G1_G1YOUNGREMSETSAMPLINGTHREAD_HPP */

hotspot/src/share/vm/gc/shared/concurrentGCThread.cpp

@@ -66,6 +66,7 @@ void ConcurrentGCThread::wait_for_universe_init() {
 }
 
 void ConcurrentGCThread::terminate() {
+  assert(_should_terminate, "Should only be called on terminate request.");
   // Signal that it is terminated
   {
     MutexLockerEx mu(Terminator_lock,