view src/share/vm/gc_implementation/concurrentMarkSweep/concurrentMarkSweepThread.hpp @ 452:00b023ae2d78

6722113: CMS: Incorrect overflow handling during precleaning of Reference lists Summary: When we encounter marking stack overflow during precleaning of Reference lists, we were using the overflow list mechanism, which can cause problems on account of mutating the mark word of the header because of conflicts with mutator accesses and updates of that field. Instead we should use the usual mechanism for overflow handling in concurrent phases, namely dirtying of the card on which the overflowed object lies. Since precleaning effectively does a form of discovered list processing, albeit with discovery enabled, we needed to adjust some code to be correct in the face of interleaved processing and discovery. Reviewed-by: apetrusenko, jcoomes
author ysr
date Thu, 20 Nov 2008 12:27:41 -0800
parents a61af66fc99e
children c18cbe5936b8
line wrap: on
line source

/*
 * Copyright 2001-2006 Sun Microsystems, Inc.  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 Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
 * CA 95054 USA or visit www.sun.com if you need additional information or
 * have any questions.
 *
 */

class ConcurrentMarkSweepGeneration;
class CMSCollector;

// The Concurrent Mark Sweep GC Thread (could be several in the future).
class ConcurrentMarkSweepThread: public ConcurrentGCThread {
  friend class VMStructs;
  friend class ConcurrentMarkSweepGeneration;   // XXX should remove friendship
  friend class CMSCollector;
 public:
  virtual void run();

 private:
  static ConcurrentMarkSweepThread*     _cmst;
  static CMSCollector*                  _collector;
  static SurrogateLockerThread*         _slt;
  static SurrogateLockerThread::SLT_msg_type _sltBuffer;
  static Monitor*                       _sltMonitor;

  ConcurrentMarkSweepThread*            _next;

  static bool _should_terminate;

  enum CMS_flag_type {
    CMS_nil             = NoBits,
    CMS_cms_wants_token = nth_bit(0),
    CMS_cms_has_token   = nth_bit(1),
    CMS_vm_wants_token  = nth_bit(2),
    CMS_vm_has_token    = nth_bit(3)
  };

  static int _CMS_flag;

  static bool CMS_flag_is_set(int b)        { return (_CMS_flag & b) != 0;   }
  static bool set_CMS_flag(int b)           { return (_CMS_flag |= b) != 0;  }
  static bool clear_CMS_flag(int b)         { return (_CMS_flag &= ~b) != 0; }
  void sleepBeforeNextCycle();

  // CMS thread should yield for a young gen collection, direct allocation,
  // and iCMS activity.
  static char _pad_1[64 - sizeof(jint)];    // prevent cache-line sharing
  static volatile jint _pending_yields;
  static volatile jint _pending_decrements; // decrements to _pending_yields
  static char _pad_2[64 - sizeof(jint)];    // prevent cache-line sharing

  // Tracing messages, enabled by CMSTraceThreadState.
  static inline void trace_state(const char* desc);

  static volatile bool _icms_enabled;   // iCMS enabled?
  static volatile bool _should_run;     // iCMS may run
  static volatile bool _should_stop;    // iCMS should stop

  // debugging
  void verify_ok_to_terminate() const PRODUCT_RETURN;

 public:
  // Constructor
  ConcurrentMarkSweepThread(CMSCollector* collector);

  static void makeSurrogateLockerThread(TRAPS);
  static SurrogateLockerThread* slt() { return _slt; }

  // Tester
  bool is_ConcurrentGC_thread() const { return true;       }

  static void threads_do(ThreadClosure* tc);

  // Printing
  void print_on(outputStream* st) const;
  void print() const                                  { print_on(tty); }
  static void print_all_on(outputStream* st);
  static void print_all()                             { print_all_on(tty); }

  // Returns the CMS Thread
  static ConcurrentMarkSweepThread* cmst()    { return _cmst; }
  static CMSCollector*         collector()    { return _collector;  }

  // Create and start the CMS Thread, or stop it on shutdown
  static ConcurrentMarkSweepThread* start(CMSCollector* collector);
  static void stop();
  static bool should_terminate() { return _should_terminate; }

  // Synchronization using CMS token
  static void synchronize(bool is_cms_thread);
  static void desynchronize(bool is_cms_thread);
  static bool vm_thread_has_cms_token() {
    return CMS_flag_is_set(CMS_vm_has_token);
  }
  static bool cms_thread_has_cms_token() {
    return CMS_flag_is_set(CMS_cms_has_token);
  }
  static bool vm_thread_wants_cms_token() {
    return CMS_flag_is_set(CMS_vm_wants_token);
  }
  static bool cms_thread_wants_cms_token() {
    return CMS_flag_is_set(CMS_cms_wants_token);
  }

  // Wait on CMS lock until the next synchronous GC
  // or given timeout, whichever is earlier.
  void    wait_on_cms_lock(long t); // milliseconds

  // The CMS thread will yield during the work portion of it's cycle
  // only when requested to.  Both synchronous and asychronous requests
  // are provided.  A synchronous request is used for young gen
  // collections and direct allocations.  The requesting thread increments
  // pending_yields at the beginning of an operation, and decrements it when
  // the operation is completed.  The CMS thread yields when pending_yields
  // is positive.  An asynchronous request is used by iCMS in the stop_icms()
  // operation. A single yield satisfies the outstanding asynch yield requests.
  // The requesting thread increments both pending_yields and pending_decrements.
  // After yielding, the CMS thread decrements both by the amount in
  // pending_decrements.
  // Note that, while "_pending_yields >= _pending_decrements" is an invariant,
  // we cannot easily test that invariant, since the counters are manipulated via
  // atomic instructions without explicit locking and we cannot read
  // the two counters atomically together: one suggestion is to
  // use (for example) 16-bit counters so as to be able to read the
  // two counters atomically even on 32-bit platforms. Notice that
  // the second assert in acknowledge_yield_request() does indeed
  // check a form of the above invariant, albeit indirectly.

  static void increment_pending_yields()   {
    Atomic::inc(&_pending_yields);
    assert(_pending_yields >= 0, "can't be negative");
  }
  static void decrement_pending_yields()   {
    Atomic::dec(&_pending_yields);
    assert(_pending_yields >= 0, "can't be negative");
  }
  static void asynchronous_yield_request() {
    increment_pending_yields();
    Atomic::inc(&_pending_decrements);
    assert(_pending_decrements >= 0, "can't be negative");
  }
  static void acknowledge_yield_request() {
    jint decrement = _pending_decrements;
    if (decrement > 0) {
      // Order important to preserve: _pending_yields >= _pending_decrements
      Atomic::add(-decrement, &_pending_decrements);
      Atomic::add(-decrement, &_pending_yields);
      assert(_pending_decrements >= 0, "can't be negative");
      assert(_pending_yields >= 0, "can't be negative");
    }
  }
  static bool should_yield()   { return _pending_yields > 0; }

  // CMS incremental mode.
  static void start_icms(); // notify thread to start a quantum of work
  static void stop_icms();  // request thread to stop working
  void icms_wait();         // if asked to stop, wait until notified to start

  // Incremental mode is enabled globally by the flag CMSIncrementalMode.  It
  // must also be enabled/disabled dynamically to allow foreground collections.
  static inline void enable_icms()              { _icms_enabled = true; }
  static inline void disable_icms()             { _icms_enabled = false; }
  static inline void set_icms_enabled(bool val) { _icms_enabled = val; }
  static inline bool icms_enabled()             { return _icms_enabled; }
};

inline void ConcurrentMarkSweepThread::trace_state(const char* desc) {
  if (CMSTraceThreadState) {
    char buf[128];
    TimeStamp& ts = gclog_or_tty->time_stamp();
    if (!ts.is_updated()) {
      ts.update();
    }
    jio_snprintf(buf, sizeof(buf), " [%.3f:  CMSThread %s] ",
                 ts.seconds(), desc);
    buf[sizeof(buf) - 1] = '\0';
    gclog_or_tty->print(buf);
  }
}

// For scoped increment/decrement of yield requests
class CMSSynchronousYieldRequest: public StackObj {
 public:
  CMSSynchronousYieldRequest() {
    ConcurrentMarkSweepThread::increment_pending_yields();
  }
  ~CMSSynchronousYieldRequest() {
    ConcurrentMarkSweepThread::decrement_pending_yields();
  }
};

// Used to emit a warning in case of unexpectedly excessive
// looping (in "apparently endless loops") in CMS code.
class CMSLoopCountWarn: public StackObj {
 private:
  const char* _src;
  const char* _msg;
  const intx  _threshold;
  intx        _ticks;

 public:
  inline CMSLoopCountWarn(const char* src, const char* msg,
                          const intx threshold) :
    _src(src), _msg(msg), _threshold(threshold), _ticks(0) { }

  inline void tick() {
    _ticks++;
    if (CMSLoopWarn && _ticks % _threshold == 0) {
      warning("%s has looped %d times %s", _src, _ticks, _msg);
    }
  }
};