Parallelize marking phase of mark-sweep/compact collection cycle.

src/atomic.h

+// Copyright 2010 the V8 project authors. All rights reserved.
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are
+// met:
+//
+//     * Redistributions of source code must retain the above copyright
+//       notice, this list of conditions and the following disclaimer.
+//     * Redistributions in binary form must reproduce the above
+//       copyright notice, this list of conditions and the following
+//       disclaimer in the documentation and/or other materials provided
+//       with the distribution.
+//     * Neither the name of Google Inc. nor the names of its
+//       contributors may be used to endorse or promote products derived
+//       from this software without specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+// This module wraps compiler specific syncronisation related intrinsics.
+
+#ifndef V8_ATOMIC_H_
+#define V8_ATOMIC_H_
+
+// Avoid warning when compiled with /Wp64.
+#ifndef _MSC_VER
+#define __w64
+#endif
+typedef __w64 int32_t Atomic32;
+#ifdef V8_TARGET_ARCH_X64
+// We need to be able to go between Atomic64 and AtomicWord implicitly.  This
+// means Atomic64 and AtomicWord should be the same type on 64-bit.
+typedef intptr_t Atomic64;
+#endif
+
+// Use AtomicWord for a machine-sized pointer.  It will use the Atomic32 or
+// Atomic64 routines below, depending on your architecture.
+typedef intptr_t AtomicWord;
+
+inline void AtomicAdd(volatile Atomic32* ptr, Atomic32 value);
+inline void AtomicOr(volatile Atomic32* ptr, Atomic32 value);
+inline void AtomicAnd(volatile Atomic32* ptr, Atomic32 value);
+inline bool AtomicCompareAndSwap(volatile Atomic32* ptr,
+                                 Atomic32 old_value,
+                                 Atomic32 new_value);
+
+#if defined(V8_TARGET_ARCH_X64)
+inline bool AtomicCompareAndSwap(volatile Atomic64* ptr,
+                                 Atomic64 old_value,
+                                 Atomic64 new_value);
+#endif
+
+
+#if defined(V8_TARGET_ARCH_IA32) || defined(V8_TARGET_ARCH_X64)
+
+// Microsoft Visual C++ specific stuff.
+#ifdef _MSC_VER
+#if (_MSC_VER >= 1500)
+#include <intrin.h>
+#else
+// For older versions we have to provide intrisic signatures.
+long _InterlockedExchangeAdd (long volatile* Addend, long Value);
+long _InterlockedOr (long volatile* Value, long Mask);
+long _InterlockedAnd (long volatile *Value, long Mask);
+long _InterlockedCompareExchange (long volatile* Destination,
+                                  long Exchange,
+                                  long Comperand);
+
+#pragma intrinsic(_InterlockedExchangeAdd)
+#pragma intrinsic(_InterlockedOr)
+#pragma intrinsic(_InterlockedAnd)
+#pragma intrinsic(_InterlockedCompareExchange)
+#endif
+
+inline void AtomicAdd(volatile Atomic32* ptr, Atomic32 value) {
+  _InterlockedExchangeAdd(reinterpret_cast<long volatile*>(ptr),
+                          static_cast<long>(value));
+}
+
+inline void AtomicOr(volatile Atomic32* ptr, Atomic32 value)  {
+  _InterlockedOr(reinterpret_cast<long volatile*>(ptr),
+                 static_cast<long>(value));
+}
+
+inline void AtomicAnd(volatile Atomic32* ptr, Atomic32 value)  {
+  _InterlockedAnd(reinterpret_cast<long volatile*>(ptr),
+                  static_cast<long>(value));
+}
+
+inline bool AtomicCompareAndSwap(volatile Atomic32* ptr,
+                                 Atomic32 old_value,
+                                 Atomic32 new_value) {
+  long result = _InterlockedCompareExchange(
+      reinterpret_cast<long volatile*>(ptr),
+      static_cast<long>(new_value),
+      static_cast<long>(old_value));
+  return result == static_cast<long>(old_value);
+}
+
+#if defined(V8_TARGET_ARCH_X64)
+inline bool AtomicCompareAndSwap(volatile Atomic64* ptr,
+                                 Atomic64 old_value,
+                                 Atomic64 new_value) {
+
+  __int64 result = _InterlockedCompareExchange_64(
+      reinterpret_cast<__int64 volatile*>(ptr),
+      static_cast<__int64>(new_value),
+      static_cast<__int64>(old_value));
+  return result == static_cast<__int64>(old_value);
+}
+#endif
+
+#define ATOMIC_SUPPORTED 1
+
+#endif  // _MSC_VER
+
+// GCC specific stuff
+#ifdef __GNUC__
+inline void AtomicAdd(volatile Atomic32* ptr, Atomic32 value) {
+  __sync_fetch_and_add(ptr, value);
+}
+
+inline void AtomicOr(volatile Atomic32* ptr, Atomic32 value)  {
+  __sync_fetch_and_or(ptr, value);
+}
+
+inline void AtomicAnd(volatile Atomic32* ptr, Atomic32 value)  {
+  __sync_fetch_and_and(ptr, value);
+}
+
+inline bool AtomicCompareAndSwap(volatile Atomic32* ptr,
+                                 Atomic32 old_value,
+                                 Atomic32 new_value) {
+  return __sync_bool_compare_and_swap(ptr, old_value, new_value);
+}
+
+#if defined(V8_TARGET_ARCH_X64)
+inline bool AtomicCompareAndSwap(volatile Atomic64* ptr,
+                                 Atomic64 old_value,
+                                 Atomic64 new_value) {
+  return __sync_bool_compare_and_swap(ptr, old_value, new_value);
+}
+#endif
+
+#define ATOMIC_SUPPORTED 1
+#endif
+
+#endif  // defined(V8_TARGET_ARCH_IA32) || defined(V8_TARGET_ARCH_X64)
+
+#ifndef ATOMIC_SUPPORTED
+inline void AtomicAdd(volatile Atomic32* ptr, Atomic32 value) {
+  *ptr += value;
+}
+
+inline void AtomicOr(volatile Atomic32* ptr, Atomic32 value)  {
+  *ptr |= value;
+}
+
+inline void AtomicAnd(volatile Atomic32* ptr, Atomic32 value)  {
+  *ptr &= value;
+}
+
+inline bool AtomicCompareAndSwap(volatile Atomic32* ptr,
+                                 Atomic32 old_value,
+                                 Atomic32 new_value) {
+  if (*ptr == old_value) {
+    *ptr = new_value;
+    return true;
+  }
+  return false;
+}
+
+#if defined(V8_TARGET_ARCH_X64)
+inline bool AtomicCompareAndSwap(volatile Atomic64* ptr,
+                                 Atomic64 old_value,
+                                 Atomic64 new_value) {
+  if (*ptr == old_value) {
+    *ptr = new_value;
+    return true;
+  }
+  return false;
+}
+#endif
+
+#define ATOMIC_SUPPORTED 0
+#endif
+
+
+#endif