Merge
authorjwilhelm
Thu, 12 Mar 2015 16:10:58 +0100
changeset 29473 2724f5121c92
parent 29472 740ba666f6bb (current diff)
parent 29463 4b16497408b9 (diff)
child 29477 82f545c6572b
Merge
hotspot/src/cpu/x86/vm/templateTable_x86_32.cpp
hotspot/src/cpu/x86/vm/templateTable_x86_64.cpp
--- a/hotspot/agent/src/share/classes/sun/jvm/hotspot/HotSpotTypeDataBase.java	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/agent/src/share/classes/sun/jvm/hotspot/HotSpotTypeDataBase.java	Thu Mar 12 16:10:58 2015 +0100
@@ -51,6 +51,9 @@
   private static final int C_INT32_SIZE = 4;
   private static final int C_INT64_SIZE = 8;
   private static int pointerSize = UNINITIALIZED_SIZE;
+  // Counter to ensure read loops terminate:
+  private static final int MAX_DUPLICATE_DEFINITIONS = 100;
+  private int duplicateDefCount = 0;
 
   private static final boolean DEBUG;
   static {
@@ -166,6 +169,10 @@
     typeEntrySizeOffset           = getLongValueFromProcess("gHotSpotVMTypeEntrySizeOffset");
     typeEntryArrayStride          = getLongValueFromProcess("gHotSpotVMTypeEntryArrayStride");
 
+    if (typeEntryArrayStride == 0L) {
+      throw new RuntimeException("zero stride: cannot read types.");
+    }
+
     // Start iterating down it until we find an entry with no name
     Address typeNameAddr = null;
     do {
@@ -192,7 +199,11 @@
       }
 
       entryAddr = entryAddr.addOffsetTo(typeEntryArrayStride);
-    } while (typeNameAddr != null);
+    } while (typeNameAddr != null && duplicateDefCount < MAX_DUPLICATE_DEFINITIONS);
+
+    if (duplicateDefCount >= MAX_DUPLICATE_DEFINITIONS) {
+      throw new RuntimeException("too many duplicate definitions");
+    }
   }
 
   private void initializePrimitiveTypes() {
@@ -395,6 +406,10 @@
     structEntryAddressOffset      = getLongValueFromProcess("gHotSpotVMStructEntryAddressOffset");
     structEntryArrayStride        = getLongValueFromProcess("gHotSpotVMStructEntryArrayStride");
 
+    if (structEntryArrayStride == 0L) {
+      throw new RuntimeException("zero stride: cannot read types.");
+    }
+
     // Fetch the address of the VMStructEntry*
     Address entryAddr = lookupInProcess("gHotSpotVMStructs");
     // Dereference this once to get the pointer to the first VMStructEntry
@@ -472,6 +487,11 @@
     intConstantEntryValueOffset = getLongValueFromProcess("gHotSpotVMIntConstantEntryValueOffset");
     intConstantEntryArrayStride = getLongValueFromProcess("gHotSpotVMIntConstantEntryArrayStride");
 
+    if (intConstantEntryArrayStride == 0L) {
+      throw new RuntimeException("zero stride: cannot read types.");
+    }
+
+
     // Fetch the address of the VMIntConstantEntry*
     Address entryAddr = lookupInProcess("gHotSpotVMIntConstants");
     // Dereference this once to get the pointer to the first VMIntConstantEntry
@@ -501,12 +521,17 @@
           } else {
             System.err.println("Warning: the int constant \"" + name + "\" (declared in the remote VM in VMStructs::localHotSpotVMIntConstants) " +
                                "had its value declared as " + value + " twice. Continuing.");
+            duplicateDefCount++;
           }
         }
       }
 
       entryAddr = entryAddr.addOffsetTo(intConstantEntryArrayStride);
-    } while (nameAddr != null);
+    } while (nameAddr != null && duplicateDefCount < MAX_DUPLICATE_DEFINITIONS);
+
+    if (duplicateDefCount >= MAX_DUPLICATE_DEFINITIONS) {
+      throw new RuntimeException("too many duplicate definitions");
+    }
   }
 
   private void readVMLongConstants() {
@@ -519,6 +544,10 @@
     longConstantEntryValueOffset = getLongValueFromProcess("gHotSpotVMLongConstantEntryValueOffset");
     longConstantEntryArrayStride = getLongValueFromProcess("gHotSpotVMLongConstantEntryArrayStride");
 
+    if (longConstantEntryArrayStride == 0L) {
+      throw new RuntimeException("zero stride: cannot read types.");
+    }
+
     // Fetch the address of the VMLongConstantEntry*
     Address entryAddr = lookupInProcess("gHotSpotVMLongConstants");
     // Dereference this once to get the pointer to the first VMLongConstantEntry
@@ -548,12 +577,17 @@
           } else {
             System.err.println("Warning: the long constant \"" + name + "\" (declared in the remote VM in VMStructs::localHotSpotVMLongConstants) " +
                                "had its value declared as " + value + " twice. Continuing.");
+            duplicateDefCount++;
           }
         }
       }
 
       entryAddr = entryAddr.addOffsetTo(longConstantEntryArrayStride);
-    } while (nameAddr != null);
+    } while (nameAddr != null && duplicateDefCount < MAX_DUPLICATE_DEFINITIONS);
+
+    if (duplicateDefCount >= MAX_DUPLICATE_DEFINITIONS) {
+      throw new RuntimeException("too many duplicate definitions.");
+    }
   }
 
   private BasicType lookupOrFail(String typeName) {
@@ -740,9 +774,10 @@
             }
 
             if (!typeNameIsPointerType(typeName)) {
-            System.err.println("Warning: the type \"" + typeName + "\" (declared in the remote VM in VMStructs::localHotSpotVMTypes) " +
-                               "had its size declared as " + size + " twice. Continuing.");
-        }
+                System.err.println("Warning: the type \"" + typeName + "\" (declared in the remote VM in VMStructs::localHotSpotVMTypes) " +
+                                   "had its size declared as " + size + " twice. Continuing.");
+                duplicateDefCount++;
+            }
         }
 
     }
--- a/hotspot/make/aix/makefiles/rules.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/aix/makefiles/rules.make	Thu Mar 12 16:10:58 2015 +0100
@@ -126,10 +126,17 @@
 RUN.JAR$(MAKE_VERBOSE) += >/dev/null
 
 # Settings for javac
-BOOT_SOURCE_LANGUAGE_VERSION = 6
-BOOT_TARGET_CLASS_VERSION = 6
 JAVAC_FLAGS = -g -encoding ascii
-BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) -source $(BOOT_SOURCE_LANGUAGE_VERSION) -target $(BOOT_TARGET_CLASS_VERSION)
+
+# Prefer BOOT_JDK_SOURCETARGET if it's set (typically by the top build system)
+# Fall back to the values here if it's not set (hotspot only builds)
+ifeq ($(BOOT_JDK_SOURCETARGET),)
+BOOTSTRAP_SOURCETARGET := -source 8 -target 8
+else
+BOOTSTRAP_SOURCETARGET := $(BOOT_JDK_SOURCETARGET)
+endif
+
+BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) $(BOOTSTRAP_SOURCETARGET)
 
 # With parallel makes, print a message at the end of compilation.
 ifeq    ($(findstring j,$(MFLAGS)),j)
--- a/hotspot/make/bsd/makefiles/rules.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/bsd/makefiles/rules.make	Thu Mar 12 16:10:58 2015 +0100
@@ -126,10 +126,17 @@
 RUN.JAR$(MAKE_VERBOSE) += >/dev/null
 
 # Settings for javac
-BOOT_SOURCE_LANGUAGE_VERSION = 6
-BOOT_TARGET_CLASS_VERSION = 6
 JAVAC_FLAGS = -g -encoding ascii
-BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) -source $(BOOT_SOURCE_LANGUAGE_VERSION) -target $(BOOT_TARGET_CLASS_VERSION)
+
+# Prefer BOOT_JDK_SOURCETARGET if it's set (typically by the top build system)
+# Fall back to the values here if it's not set (hotspot only builds)
+ifeq ($(BOOT_JDK_SOURCETARGET),)
+BOOTSTRAP_SOURCETARGET := -source 8 -target 8
+else
+BOOTSTRAP_SOURCETARGET := $(BOOT_JDK_SOURCETARGET)
+endif
+
+BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) $(BOOTSTRAP_SOURCETARGET)
 
 # With parallel makes, print a message at the end of compilation.
 ifeq    ($(findstring j,$(MFLAGS)),j)
--- a/hotspot/make/defs.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/defs.make	Thu Mar 12 16:10:58 2015 +0100
@@ -347,6 +347,8 @@
 # includes this make/defs.make file.
 MAKE_ARGS += HOTSPOT_BUILD_VERSION=$(HOTSPOT_BUILD_VERSION)
 
+MAKE_ARGS += BOOT_JDK_SOURCETARGET="$(BOOT_JDK_SOURCETARGET)"
+
 # Various export sub directories
 EXPORT_INCLUDE_DIR = $(EXPORT_PATH)/include
 EXPORT_DOCS_DIR = $(EXPORT_PATH)/docs
--- a/hotspot/make/linux/makefiles/rules.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/linux/makefiles/rules.make	Thu Mar 12 16:10:58 2015 +0100
@@ -126,10 +126,17 @@
 RUN.JAR$(MAKE_VERBOSE) += >/dev/null
 
 # Settings for javac
-BOOT_SOURCE_LANGUAGE_VERSION = 6
-BOOT_TARGET_CLASS_VERSION = 6
 JAVAC_FLAGS = -g -encoding ascii
-BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) -source $(BOOT_SOURCE_LANGUAGE_VERSION) -target $(BOOT_TARGET_CLASS_VERSION)
+
+# Prefer BOOT_JDK_SOURCETARGET if it's set (typically by the top build system)
+# Fall back to the values here if it's not set (hotspot only builds)
+ifeq ($(BOOT_JDK_SOURCETARGET),)
+BOOTSTRAP_SOURCETARGET := -source 8 -target 8
+else
+BOOTSTRAP_SOURCETARGET := $(BOOT_JDK_SOURCETARGET)
+endif
+
+BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) $(BOOTSTRAP_SOURCETARGET)
 
 # With parallel makes, print a message at the end of compilation.
 ifeq    ($(findstring j,$(MFLAGS)),j)
--- a/hotspot/make/solaris/makefiles/rules.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/solaris/makefiles/rules.make	Thu Mar 12 16:10:58 2015 +0100
@@ -118,10 +118,17 @@
 RUN.JAR$(MAKE_VERBOSE) += >/dev/null
 
 # Settings for javac
-BOOT_SOURCE_LANGUAGE_VERSION = 6
-BOOT_TARGET_CLASS_VERSION = 6
 JAVAC_FLAGS = -g -encoding ascii
-BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) -source $(BOOT_SOURCE_LANGUAGE_VERSION) -target $(BOOT_TARGET_CLASS_VERSION)
+
+# Prefer BOOT_JDK_SOURCETARGET if it's set (typically by the top build system)
+# Fall back to the values here if it's not set (hotspot only builds)
+ifeq ($(BOOT_JDK_SOURCETARGET),)
+BOOTSTRAP_SOURCETARGET := -source 8 -target 8
+else
+BOOTSTRAP_SOURCETARGET := $(BOOT_JDK_SOURCETARGET)
+endif
+
+BOOTSTRAP_JAVAC_FLAGS = $(JAVAC_FLAGS) $(BOOTSTRAP_SOURCETARGET)
 
 # With parallel makes, print a message at the end of compilation.
 ifeq    ($(findstring j,$(MFLAGS)),j)
--- a/hotspot/make/windows/makefiles/rules.make	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/make/windows/makefiles/rules.make	Thu Mar 12 16:10:58 2015 +0100
@@ -44,10 +44,17 @@
 !endif
 
 # Settings for javac
-BOOT_SOURCE_LANGUAGE_VERSION=6
-BOOT_TARGET_CLASS_VERSION=6
 JAVAC_FLAGS=-g -encoding ascii
-BOOTSTRAP_JAVAC_FLAGS=$(JAVAC_FLAGS) -source $(BOOT_SOURCE_LANGUAGE_VERSION) -target $(BOOT_TARGET_CLASS_VERSION)
+
+# Prefer BOOT_JDK_SOURCETARGET if it's set (typically by the top build system)
+# Fall back to the values here if it's not set (hotspot only builds)
+!ifndef BOOT_JDK_SOURCETARGET
+BOOTSTRAP_SOURCETARGET=-source 8 -target 8
+!else
+BOOTSTRAP_SOURCETARGET=$(BOOT_JDK_SOURCETARGET)
+!endif
+
+BOOTSTRAP_JAVAC_FLAGS=$(JAVAC_FLAGS) $(BOOTSTRAP_SOURCETARGET)
 
 # VS2012 and VS2013 loads VS10 projects just fine (and will
 # upgrade them automatically to VS2012 format).
--- a/hotspot/src/cpu/x86/vm/c1_LIRAssembler_x86.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/cpu/x86/vm/c1_LIRAssembler_x86.cpp	Thu Mar 12 16:10:58 2015 +0100
@@ -3947,12 +3947,10 @@
 
 void LIR_Assembler::membar_acquire() {
   // No x86 machines currently require load fences
-  // __ load_fence();
 }
 
 void LIR_Assembler::membar_release() {
   // No x86 machines currently require store fences
-  // __ store_fence();
 }
 
 void LIR_Assembler::membar_loadload() {
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/hotspot/src/cpu/x86/vm/templateTable_x86.cpp	Thu Mar 12 16:10:58 2015 +0100
@@ -0,0 +1,4235 @@
+/*
+ * Copyright (c) 1997, 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 "asm/macroAssembler.hpp"
+#include "interpreter/interpreter.hpp"
+#include "interpreter/interpreterRuntime.hpp"
+#include "interpreter/interp_masm.hpp"
+#include "interpreter/templateTable.hpp"
+#include "memory/universe.inline.hpp"
+#include "oops/methodData.hpp"
+#include "oops/objArrayKlass.hpp"
+#include "oops/oop.inline.hpp"
+#include "prims/methodHandles.hpp"
+#include "runtime/sharedRuntime.hpp"
+#include "runtime/stubRoutines.hpp"
+#include "runtime/synchronizer.hpp"
+#include "utilities/macros.hpp"
+
+#ifndef CC_INTERP
+
+#define __ _masm->
+
+// Global Register Names
+Register rbcp     = LP64_ONLY(r13) NOT_LP64(rsi);
+Register rlocals  = LP64_ONLY(r14) NOT_LP64(rdi);
+
+// Platform-dependent initialization
+void TemplateTable::pd_initialize() {
+  // No x86 specific initialization
+}
+
+// Address Computation: local variables
+static inline Address iaddress(int n) {
+  return Address(rlocals, Interpreter::local_offset_in_bytes(n));
+}
+
+static inline Address laddress(int n) {
+  return iaddress(n + 1);
+}
+
+#ifndef _LP64
+static inline Address haddress(int n) {
+  return iaddress(n + 0);
+}
+#endif
+
+static inline Address faddress(int n) {
+  return iaddress(n);
+}
+
+static inline Address daddress(int n) {
+  return laddress(n);
+}
+
+static inline Address aaddress(int n) {
+  return iaddress(n);
+}
+
+static inline Address iaddress(Register r) {
+  return Address(rlocals, r, Address::times_ptr);
+}
+
+static inline Address laddress(Register r) {
+  return Address(rlocals, r, Address::times_ptr, Interpreter::local_offset_in_bytes(1));
+}
+
+#ifndef _LP64
+static inline Address haddress(Register r)       {
+  return Address(rlocals, r, Interpreter::stackElementScale(), Interpreter::local_offset_in_bytes(0));
+}
+#endif
+
+static inline Address faddress(Register r) {
+  return iaddress(r);
+}
+
+static inline Address daddress(Register r) {
+  return laddress(r);
+}
+
+static inline Address aaddress(Register r) {
+  return iaddress(r);
+}
+
+
+// expression stack
+// (Note: Must not use symmetric equivalents at_rsp_m1/2 since they store
+// data beyond the rsp which is potentially unsafe in an MT environment;
+// an interrupt may overwrite that data.)
+static inline Address at_rsp   () {
+  return Address(rsp, 0);
+}
+
+// At top of Java expression stack which may be different than esp().  It
+// isn't for category 1 objects.
+static inline Address at_tos   () {
+  return Address(rsp,  Interpreter::expr_offset_in_bytes(0));
+}
+
+static inline Address at_tos_p1() {
+  return Address(rsp,  Interpreter::expr_offset_in_bytes(1));
+}
+
+static inline Address at_tos_p2() {
+  return Address(rsp,  Interpreter::expr_offset_in_bytes(2));
+}
+
+// Condition conversion
+static Assembler::Condition j_not(TemplateTable::Condition cc) {
+  switch (cc) {
+  case TemplateTable::equal        : return Assembler::notEqual;
+  case TemplateTable::not_equal    : return Assembler::equal;
+  case TemplateTable::less         : return Assembler::greaterEqual;
+  case TemplateTable::less_equal   : return Assembler::greater;
+  case TemplateTable::greater      : return Assembler::lessEqual;
+  case TemplateTable::greater_equal: return Assembler::less;
+  }
+  ShouldNotReachHere();
+  return Assembler::zero;
+}
+
+
+
+// Miscelaneous helper routines
+// Store an oop (or NULL) at the address described by obj.
+// If val == noreg this means store a NULL
+
+
+static void do_oop_store(InterpreterMacroAssembler* _masm,
+                         Address obj,
+                         Register val,
+                         BarrierSet::Name barrier,
+                         bool precise) {
+  assert(val == noreg || val == rax, "parameter is just for looks");
+  switch (barrier) {
+#if INCLUDE_ALL_GCS
+    case BarrierSet::G1SATBCT:
+    case BarrierSet::G1SATBCTLogging:
+      {
+        // flatten object address if needed
+        // We do it regardless of precise because we need the registers
+        if (obj.index() == noreg && obj.disp() == 0) {
+          if (obj.base() != rdx) {
+            __ movptr(rdx, obj.base());
+          }
+        } else {
+          __ lea(rdx, obj);
+        }
+
+        Register rtmp    = LP64_ONLY(r8)         NOT_LP64(rsi);
+        Register rthread = LP64_ONLY(r15_thread) NOT_LP64(rcx);
+
+        NOT_LP64(__ get_thread(rcx));
+        NOT_LP64(__ save_bcp());
+
+        __ g1_write_barrier_pre(rdx /* obj */,
+                                rbx /* pre_val */,
+                                rthread /* thread */,
+                                rtmp  /* tmp */,
+                                val != noreg /* tosca_live */,
+                                false /* expand_call */);
+        if (val == noreg) {
+          __ store_heap_oop_null(Address(rdx, 0));
+        } else {
+          // G1 barrier needs uncompressed oop for region cross check.
+          Register new_val = val;
+          if (UseCompressedOops) {
+            new_val = rbx;
+            __ movptr(new_val, val);
+          }
+          __ store_heap_oop(Address(rdx, 0), val);
+          __ g1_write_barrier_post(rdx /* store_adr */,
+                                   new_val /* new_val */,
+                                   rthread /* thread */,
+                                   rtmp /* tmp */,
+                                   rbx /* tmp2 */);
+        }
+        NOT_LP64( __ restore_bcp());
+      }
+      break;
+#endif // INCLUDE_ALL_GCS
+    case BarrierSet::CardTableModRef:
+    case BarrierSet::CardTableExtension:
+      {
+        if (val == noreg) {
+          __ store_heap_oop_null(obj);
+        } else {
+          __ store_heap_oop(obj, val);
+          // flatten object address if needed
+          if (!precise || (obj.index() == noreg && obj.disp() == 0)) {
+            __ store_check(obj.base());
+          } else {
+            __ lea(rdx, obj);
+            __ store_check(rdx);
+          }
+        }
+      }
+      break;
+    case BarrierSet::ModRef:
+      if (val == noreg) {
+        __ store_heap_oop_null(obj);
+      } else {
+        __ store_heap_oop(obj, val);
+      }
+      break;
+    default      :
+      ShouldNotReachHere();
+
+  }
+}
+
+Address TemplateTable::at_bcp(int offset) {
+  assert(_desc->uses_bcp(), "inconsistent uses_bcp information");
+  return Address(rbcp, offset);
+}
+
+
+void TemplateTable::patch_bytecode(Bytecodes::Code bc, Register bc_reg,
+                                   Register temp_reg, bool load_bc_into_bc_reg/*=true*/,
+                                   int byte_no) {
+  if (!RewriteBytecodes)  return;
+  Label L_patch_done;
+
+  switch (bc) {
+  case Bytecodes::_fast_aputfield:
+  case Bytecodes::_fast_bputfield:
+  case Bytecodes::_fast_cputfield:
+  case Bytecodes::_fast_dputfield:
+  case Bytecodes::_fast_fputfield:
+  case Bytecodes::_fast_iputfield:
+  case Bytecodes::_fast_lputfield:
+  case Bytecodes::_fast_sputfield:
+    {
+      // We skip bytecode quickening for putfield instructions when
+      // the put_code written to the constant pool cache is zero.
+      // This is required so that every execution of this instruction
+      // calls out to InterpreterRuntime::resolve_get_put to do
+      // additional, required work.
+      assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
+      assert(load_bc_into_bc_reg, "we use bc_reg as temp");
+      __ get_cache_and_index_and_bytecode_at_bcp(temp_reg, bc_reg, temp_reg, byte_no, 1);
+      __ movl(bc_reg, bc);
+      __ cmpl(temp_reg, (int) 0);
+      __ jcc(Assembler::zero, L_patch_done);  // don't patch
+    }
+    break;
+  default:
+    assert(byte_no == -1, "sanity");
+    // the pair bytecodes have already done the load.
+    if (load_bc_into_bc_reg) {
+      __ movl(bc_reg, bc);
+    }
+  }
+
+  if (JvmtiExport::can_post_breakpoint()) {
+    Label L_fast_patch;
+    // if a breakpoint is present we can't rewrite the stream directly
+    __ movzbl(temp_reg, at_bcp(0));
+    __ cmpl(temp_reg, Bytecodes::_breakpoint);
+    __ jcc(Assembler::notEqual, L_fast_patch);
+    __ get_method(temp_reg);
+    // Let breakpoint table handling rewrite to quicker bytecode
+    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::set_original_bytecode_at), temp_reg, rbcp, bc_reg);
+#ifndef ASSERT
+    __ jmpb(L_patch_done);
+#else
+    __ jmp(L_patch_done);
+#endif
+    __ bind(L_fast_patch);
+  }
+
+#ifdef ASSERT
+  Label L_okay;
+  __ load_unsigned_byte(temp_reg, at_bcp(0));
+  __ cmpl(temp_reg, (int) Bytecodes::java_code(bc));
+  __ jcc(Assembler::equal, L_okay);
+  __ cmpl(temp_reg, bc_reg);
+  __ jcc(Assembler::equal, L_okay);
+  __ stop("patching the wrong bytecode");
+  __ bind(L_okay);
+#endif
+
+  // patch bytecode
+  __ movb(at_bcp(0), bc_reg);
+  __ bind(L_patch_done);
+}
+// Individual instructions
+
+
+void TemplateTable::nop() {
+  transition(vtos, vtos);
+  // nothing to do
+}
+
+void TemplateTable::shouldnotreachhere() {
+  transition(vtos, vtos);
+  __ stop("shouldnotreachhere bytecode");
+}
+
+void TemplateTable::aconst_null() {
+  transition(vtos, atos);
+  __ xorl(rax, rax);
+}
+
+void TemplateTable::iconst(int value) {
+  transition(vtos, itos);
+  if (value == 0) {
+    __ xorl(rax, rax);
+  } else {
+    __ movl(rax, value);
+  }
+}
+
+void TemplateTable::lconst(int value) {
+  transition(vtos, ltos);
+  if (value == 0) {
+    __ xorl(rax, rax);
+  } else {
+    __ movl(rax, value);
+  }
+#ifndef _LP64
+  assert(value >= 0, "check this code");
+  __ xorptr(rdx, rdx);
+#endif
+}
+
+
+
+void TemplateTable::fconst(int value) {
+  transition(vtos, ftos);
+#ifdef _LP64
+  static float one = 1.0f, two = 2.0f;
+  switch (value) {
+  case 0:
+    __ xorps(xmm0, xmm0);
+    break;
+  case 1:
+    __ movflt(xmm0, ExternalAddress((address) &one));
+    break;
+  case 2:
+    __ movflt(xmm0, ExternalAddress((address) &two));
+    break;
+  default:
+    ShouldNotReachHere();
+    break;
+  }
+#else
+         if (value == 0) { __ fldz();
+  } else if (value == 1) { __ fld1();
+  } else if (value == 2) { __ fld1(); __ fld1(); __ faddp(); // should do a better solution here
+  } else                 { ShouldNotReachHere();
+  }
+#endif
+}
+
+void TemplateTable::dconst(int value) {
+  transition(vtos, dtos);
+#ifdef _LP64
+  static double one = 1.0;
+  switch (value) {
+  case 0:
+    __ xorpd(xmm0, xmm0);
+    break;
+  case 1:
+    __ movdbl(xmm0, ExternalAddress((address) &one));
+    break;
+  default:
+    ShouldNotReachHere();
+    break;
+  }
+
+#else
+         if (value == 0) { __ fldz();
+  } else if (value == 1) { __ fld1();
+  } else                 { ShouldNotReachHere();
+  }
+#endif
+}
+
+void TemplateTable::bipush() {
+  transition(vtos, itos);
+  __ load_signed_byte(rax, at_bcp(1));
+}
+
+void TemplateTable::sipush() {
+  transition(vtos, itos);
+  __ load_unsigned_short(rax, at_bcp(1));
+  __ bswapl(rax);
+  __ sarl(rax, 16);
+}
+
+void TemplateTable::ldc(bool wide) {
+  transition(vtos, vtos);
+  Register rarg = NOT_LP64(rcx) LP64_ONLY(c_rarg1);
+  Label call_ldc, notFloat, notClass, Done;
+
+  if (wide) {
+    __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
+  } else {
+    __ load_unsigned_byte(rbx, at_bcp(1));
+  }
+
+  __ get_cpool_and_tags(rcx, rax);
+  const int base_offset = ConstantPool::header_size() * wordSize;
+  const int tags_offset = Array<u1>::base_offset_in_bytes();
+
+  // get type
+  __ movzbl(rdx, Address(rax, rbx, Address::times_1, tags_offset));
+
+  // unresolved class - get the resolved class
+  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClass);
+  __ jccb(Assembler::equal, call_ldc);
+
+  // unresolved class in error state - call into runtime to throw the error
+  // from the first resolution attempt
+  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClassInError);
+  __ jccb(Assembler::equal, call_ldc);
+
+  // resolved class - need to call vm to get java mirror of the class
+  __ cmpl(rdx, JVM_CONSTANT_Class);
+  __ jcc(Assembler::notEqual, notClass);
+
+  __ bind(call_ldc);
+
+  __ movl(rarg, wide);
+  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::ldc), rarg);
+
+  __ push(atos);
+  __ jmp(Done);
+
+  __ bind(notClass);
+  __ cmpl(rdx, JVM_CONSTANT_Float);
+  __ jccb(Assembler::notEqual, notFloat);
+
+  // ftos
+  LP64_ONLY(__ movflt(xmm0, Address(rcx, rbx, Address::times_8, base_offset)));
+  NOT_LP64(__ fld_s(    Address(rcx, rbx, Address::times_ptr, base_offset)));
+  __ push(ftos);
+  __ jmp(Done);
+
+  __ bind(notFloat);
+#ifdef ASSERT
+  {
+    Label L;
+    __ cmpl(rdx, JVM_CONSTANT_Integer);
+    __ jcc(Assembler::equal, L);
+    // String and Object are rewritten to fast_aldc
+    __ stop("unexpected tag type in ldc");
+    __ bind(L);
+  }
+#endif
+  // itos JVM_CONSTANT_Integer only
+  __ movl(rax, Address(rcx, rbx, Address::times_ptr, base_offset));
+  __ push(itos);
+  __ bind(Done);
+}
+
+// Fast path for caching oop constants.
+void TemplateTable::fast_aldc(bool wide) {
+  transition(vtos, atos);
+
+  Register result = rax;
+  Register tmp = rdx;
+  int index_size = wide ? sizeof(u2) : sizeof(u1);
+
+  Label resolved;
+
+  // We are resolved if the resolved reference cache entry contains a
+  // non-null object (String, MethodType, etc.)
+  assert_different_registers(result, tmp);
+  __ get_cache_index_at_bcp(tmp, 1, index_size);
+  __ load_resolved_reference_at_index(result, tmp);
+  __ testl(result, result);
+  __ jcc(Assembler::notZero, resolved);
+
+  address entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_ldc);
+
+  // first time invocation - must resolve first
+  __ movl(tmp, (int)bytecode());
+  __ call_VM(result, entry, tmp);
+
+  __ bind(resolved);
+
+  if (VerifyOops) {
+    __ verify_oop(result);
+  }
+}
+
+void TemplateTable::ldc2_w() {
+  transition(vtos, vtos);
+  Label Long, Done;
+  __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
+
+  __ get_cpool_and_tags(rcx, rax);
+  const int base_offset = ConstantPool::header_size() * wordSize;
+  const int tags_offset = Array<u1>::base_offset_in_bytes();
+
+  // get type
+  __ cmpb(Address(rax, rbx, Address::times_1, tags_offset),
+          JVM_CONSTANT_Double);
+  __ jccb(Assembler::notEqual, Long);
+
+  // dtos
+  LP64_ONLY(__ movdbl(xmm0, Address(rcx, rbx, Address::times_8, base_offset)));
+  NOT_LP64(__ fld_d(    Address(rcx, rbx, Address::times_ptr, base_offset)));
+  __ push(dtos);
+
+  __ jmpb(Done);
+  __ bind(Long);
+
+  // ltos
+  __ movptr(rax, Address(rcx, rbx, Address::times_ptr, base_offset + 0 * wordSize));
+  NOT_LP64(__ movptr(rdx, Address(rcx, rbx, Address::times_ptr, base_offset + 1 * wordSize)));
+  __ push(ltos);
+
+  __ bind(Done);
+}
+
+void TemplateTable::locals_index(Register reg, int offset) {
+  __ load_unsigned_byte(reg, at_bcp(offset));
+  __ negptr(reg);
+}
+
+void TemplateTable::iload() {
+  transition(vtos, itos);
+  if (RewriteFrequentPairs) {
+    Label rewrite, done;
+    const Register bc = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+    LP64_ONLY(assert(rbx != bc, "register damaged"));
+
+    // get next byte
+    __ load_unsigned_byte(rbx,
+                          at_bcp(Bytecodes::length_for(Bytecodes::_iload)));
+    // if _iload, wait to rewrite to iload2.  We only want to rewrite the
+    // last two iloads in a pair.  Comparing against fast_iload means that
+    // the next bytecode is neither an iload or a caload, and therefore
+    // an iload pair.
+    __ cmpl(rbx, Bytecodes::_iload);
+    __ jcc(Assembler::equal, done);
+
+    __ cmpl(rbx, Bytecodes::_fast_iload);
+    __ movl(bc, Bytecodes::_fast_iload2);
+
+    __ jccb(Assembler::equal, rewrite);
+
+    // if _caload, rewrite to fast_icaload
+    __ cmpl(rbx, Bytecodes::_caload);
+    __ movl(bc, Bytecodes::_fast_icaload);
+    __ jccb(Assembler::equal, rewrite);
+
+    // rewrite so iload doesn't check again.
+    __ movl(bc, Bytecodes::_fast_iload);
+
+    // rewrite
+    // bc: fast bytecode
+    __ bind(rewrite);
+    patch_bytecode(Bytecodes::_iload, bc, rbx, false);
+    __ bind(done);
+  }
+
+  // Get the local value into tos
+  locals_index(rbx);
+  __ movl(rax, iaddress(rbx));
+}
+
+void TemplateTable::fast_iload2() {
+  transition(vtos, itos);
+  locals_index(rbx);
+  __ movl(rax, iaddress(rbx));
+  __ push(itos);
+  locals_index(rbx, 3);
+  __ movl(rax, iaddress(rbx));
+}
+
+void TemplateTable::fast_iload() {
+  transition(vtos, itos);
+  locals_index(rbx);
+  __ movl(rax, iaddress(rbx));
+}
+
+void TemplateTable::lload() {
+  transition(vtos, ltos);
+  locals_index(rbx);
+  __ movptr(rax, laddress(rbx));
+  NOT_LP64(__ movl(rdx, haddress(rbx)));
+}
+
+void TemplateTable::fload() {
+  transition(vtos, ftos);
+  locals_index(rbx);
+  LP64_ONLY(__ movflt(xmm0, faddress(rbx)));
+  NOT_LP64(__ fld_s(faddress(rbx)));
+}
+
+void TemplateTable::dload() {
+  transition(vtos, dtos);
+  locals_index(rbx);
+  LP64_ONLY(__ movdbl(xmm0, daddress(rbx)));
+  NOT_LP64(__ fld_d(daddress(rbx)));
+}
+
+void TemplateTable::aload() {
+  transition(vtos, atos);
+  locals_index(rbx);
+  __ movptr(rax, aaddress(rbx));
+}
+
+void TemplateTable::locals_index_wide(Register reg) {
+  __ load_unsigned_short(reg, at_bcp(2));
+  __ bswapl(reg);
+  __ shrl(reg, 16);
+  __ negptr(reg);
+}
+
+void TemplateTable::wide_iload() {
+  transition(vtos, itos);
+  locals_index_wide(rbx);
+  __ movl(rax, iaddress(rbx));
+}
+
+void TemplateTable::wide_lload() {
+  transition(vtos, ltos);
+  locals_index_wide(rbx);
+  __ movptr(rax, laddress(rbx));
+  NOT_LP64(__ movl(rdx, haddress(rbx)));
+}
+
+void TemplateTable::wide_fload() {
+  transition(vtos, ftos);
+  locals_index_wide(rbx);
+  LP64_ONLY(__ movflt(xmm0, faddress(rbx)));
+  NOT_LP64(__ fld_s(faddress(rbx)));
+}
+
+void TemplateTable::wide_dload() {
+  transition(vtos, dtos);
+  locals_index_wide(rbx);
+  LP64_ONLY(__ movdbl(xmm0, daddress(rbx)));
+  NOT_LP64(__ fld_d(daddress(rbx)));
+}
+
+void TemplateTable::wide_aload() {
+  transition(vtos, atos);
+  locals_index_wide(rbx);
+  __ movptr(rax, aaddress(rbx));
+}
+
+void TemplateTable::index_check(Register array, Register index) {
+  // Pop ptr into array
+  __ pop_ptr(array);
+  index_check_without_pop(array, index);
+}
+
+void TemplateTable::index_check_without_pop(Register array, Register index) {
+  // destroys rbx
+  // check array
+  __ null_check(array, arrayOopDesc::length_offset_in_bytes());
+  // sign extend index for use by indexed load
+  __ movl2ptr(index, index);
+  // check index
+  __ cmpl(index, Address(array, arrayOopDesc::length_offset_in_bytes()));
+  if (index != rbx) {
+    // ??? convention: move aberrant index into rbx for exception message
+    assert(rbx != array, "different registers");
+    __ movl(rbx, index);
+  }
+  __ jump_cc(Assembler::aboveEqual,
+             ExternalAddress(Interpreter::_throw_ArrayIndexOutOfBoundsException_entry));
+}
+
+
+void TemplateTable::iaload() {
+  transition(itos, itos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ movl(rax, Address(rdx, rax,
+                       Address::times_4,
+                       arrayOopDesc::base_offset_in_bytes(T_INT)));
+}
+
+void TemplateTable::laload() {
+  transition(itos, ltos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  NOT_LP64(__ mov(rbx, rax));
+  // rbx,: index
+  __ movptr(rax, Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 0 * wordSize));
+  NOT_LP64(__ movl(rdx, Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 1 * wordSize)));
+}
+
+
+
+void TemplateTable::faload() {
+  transition(itos, ftos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  LP64_ONLY(__ movflt(xmm0, Address(rdx, rax,
+                         Address::times_4,
+                         arrayOopDesc::base_offset_in_bytes(T_FLOAT))));
+  NOT_LP64(__ fld_s(Address(rdx, rax, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_FLOAT))));
+}
+
+void TemplateTable::daload() {
+  transition(itos, dtos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  LP64_ONLY(__ movdbl(xmm0, Address(rdx, rax,
+                          Address::times_8,
+                          arrayOopDesc::base_offset_in_bytes(T_DOUBLE))));
+  NOT_LP64(__ fld_d(Address(rdx, rax, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_DOUBLE))));
+}
+
+void TemplateTable::aaload() {
+  transition(itos, atos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ load_heap_oop(rax, Address(rdx, rax,
+                                UseCompressedOops ? Address::times_4 : Address::times_ptr,
+                                arrayOopDesc::base_offset_in_bytes(T_OBJECT)));
+}
+
+void TemplateTable::baload() {
+  transition(itos, itos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ load_signed_byte(rax, Address(rdx, rax, Address::times_1, arrayOopDesc::base_offset_in_bytes(T_BYTE)));
+}
+
+void TemplateTable::caload() {
+  transition(itos, itos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ load_unsigned_short(rax, Address(rdx, rax, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_CHAR)));
+}
+
+// iload followed by caload frequent pair
+void TemplateTable::fast_icaload() {
+  transition(vtos, itos);
+  // load index out of locals
+  locals_index(rbx);
+  __ movl(rax, iaddress(rbx));
+
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ load_unsigned_short(rax,
+                         Address(rdx, rax,
+                                 Address::times_2,
+                                 arrayOopDesc::base_offset_in_bytes(T_CHAR)));
+}
+
+
+void TemplateTable::saload() {
+  transition(itos, itos);
+  // rax: index
+  // rdx: array
+  index_check(rdx, rax); // kills rbx
+  __ load_signed_short(rax, Address(rdx, rax, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_SHORT)));
+}
+
+void TemplateTable::iload(int n) {
+  transition(vtos, itos);
+  __ movl(rax, iaddress(n));
+}
+
+void TemplateTable::lload(int n) {
+  transition(vtos, ltos);
+  __ movptr(rax, laddress(n));
+  NOT_LP64(__ movptr(rdx, haddress(n)));
+}
+
+void TemplateTable::fload(int n) {
+  transition(vtos, ftos);
+  LP64_ONLY(__ movflt(xmm0, faddress(n)));
+  NOT_LP64(__ fld_s(faddress(n)));
+}
+
+void TemplateTable::dload(int n) {
+  transition(vtos, dtos);
+  LP64_ONLY(__ movdbl(xmm0, daddress(n)));
+  NOT_LP64(__ fld_d(daddress(n)));
+}
+
+void TemplateTable::aload(int n) {
+  transition(vtos, atos);
+  __ movptr(rax, aaddress(n));
+}
+
+void TemplateTable::aload_0() {
+  transition(vtos, atos);
+  // According to bytecode histograms, the pairs:
+  //
+  // _aload_0, _fast_igetfield
+  // _aload_0, _fast_agetfield
+  // _aload_0, _fast_fgetfield
+  //
+  // occur frequently. If RewriteFrequentPairs is set, the (slow)
+  // _aload_0 bytecode checks if the next bytecode is either
+  // _fast_igetfield, _fast_agetfield or _fast_fgetfield and then
+  // rewrites the current bytecode into a pair bytecode; otherwise it
+  // rewrites the current bytecode into _fast_aload_0 that doesn't do
+  // the pair check anymore.
+  //
+  // Note: If the next bytecode is _getfield, the rewrite must be
+  //       delayed, otherwise we may miss an opportunity for a pair.
+  //
+  // Also rewrite frequent pairs
+  //   aload_0, aload_1
+  //   aload_0, iload_1
+  // These bytecodes with a small amount of code are most profitable
+  // to rewrite
+  if (RewriteFrequentPairs) {
+    Label rewrite, done;
+
+    const Register bc = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+    LP64_ONLY(assert(rbx != bc, "register damaged"));
+
+    // get next byte
+    __ load_unsigned_byte(rbx, at_bcp(Bytecodes::length_for(Bytecodes::_aload_0)));
+
+    // do actual aload_0
+    aload(0);
+
+    // if _getfield then wait with rewrite
+    __ cmpl(rbx, Bytecodes::_getfield);
+    __ jcc(Assembler::equal, done);
+
+    // if _igetfield then reqrite to _fast_iaccess_0
+    assert(Bytecodes::java_code(Bytecodes::_fast_iaccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
+    __ cmpl(rbx, Bytecodes::_fast_igetfield);
+    __ movl(bc, Bytecodes::_fast_iaccess_0);
+    __ jccb(Assembler::equal, rewrite);
+
+    // if _agetfield then reqrite to _fast_aaccess_0
+    assert(Bytecodes::java_code(Bytecodes::_fast_aaccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
+    __ cmpl(rbx, Bytecodes::_fast_agetfield);
+    __ movl(bc, Bytecodes::_fast_aaccess_0);
+    __ jccb(Assembler::equal, rewrite);
+
+    // if _fgetfield then reqrite to _fast_faccess_0
+    assert(Bytecodes::java_code(Bytecodes::_fast_faccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
+    __ cmpl(rbx, Bytecodes::_fast_fgetfield);
+    __ movl(bc, Bytecodes::_fast_faccess_0);
+    __ jccb(Assembler::equal, rewrite);
+
+    // else rewrite to _fast_aload0
+    assert(Bytecodes::java_code(Bytecodes::_fast_aload_0) == Bytecodes::_aload_0, "fix bytecode definition");
+    __ movl(bc, Bytecodes::_fast_aload_0);
+
+    // rewrite
+    // bc: fast bytecode
+    __ bind(rewrite);
+    patch_bytecode(Bytecodes::_aload_0, bc, rbx, false);
+
+    __ bind(done);
+  } else {
+    aload(0);
+  }
+}
+
+void TemplateTable::istore() {
+  transition(itos, vtos);
+  locals_index(rbx);
+  __ movl(iaddress(rbx), rax);
+}
+
+
+void TemplateTable::lstore() {
+  transition(ltos, vtos);
+  locals_index(rbx);
+  __ movptr(laddress(rbx), rax);
+  NOT_LP64(__ movptr(haddress(rbx), rdx));
+}
+
+void TemplateTable::fstore() {
+  transition(ftos, vtos);
+  locals_index(rbx);
+  LP64_ONLY(__ movflt(faddress(rbx), xmm0));
+  NOT_LP64(__ fstp_s(faddress(rbx)));
+}
+
+void TemplateTable::dstore() {
+  transition(dtos, vtos);
+  locals_index(rbx);
+  LP64_ONLY(__ movdbl(daddress(rbx), xmm0));
+  NOT_LP64(__ fstp_d(daddress(rbx)));
+}
+
+void TemplateTable::astore() {
+  transition(vtos, vtos);
+  __ pop_ptr(rax);
+  locals_index(rbx);
+  __ movptr(aaddress(rbx), rax);
+}
+
+void TemplateTable::wide_istore() {
+  transition(vtos, vtos);
+  __ pop_i();
+  locals_index_wide(rbx);
+  __ movl(iaddress(rbx), rax);
+}
+
+void TemplateTable::wide_lstore() {
+  transition(vtos, vtos);
+  NOT_LP64(__ pop_l(rax, rdx));
+  LP64_ONLY(__ pop_l());
+  locals_index_wide(rbx);
+  __ movptr(laddress(rbx), rax);
+  NOT_LP64(__ movl(haddress(rbx), rdx));
+}
+
+void TemplateTable::wide_fstore() {
+#ifdef _LP64
+  transition(vtos, vtos);
+  __ pop_f();
+  locals_index_wide(rbx);
+  __ movflt(faddress(rbx), xmm0);
+#else
+  wide_istore();
+#endif
+}
+
+void TemplateTable::wide_dstore() {
+#ifdef _LP64
+  transition(vtos, vtos);
+  __ pop_d();
+  locals_index_wide(rbx);
+  __ movdbl(daddress(rbx), xmm0);
+#else
+  wide_lstore();
+#endif
+}
+
+void TemplateTable::wide_astore() {
+  transition(vtos, vtos);
+  __ pop_ptr(rax);
+  locals_index_wide(rbx);
+  __ movptr(aaddress(rbx), rax);
+}
+
+void TemplateTable::iastore() {
+  transition(itos, vtos);
+  __ pop_i(rbx);
+  // rax: value
+  // rbx: index
+  // rdx: array
+  index_check(rdx, rbx); // prefer index in rbx
+  __ movl(Address(rdx, rbx,
+                  Address::times_4,
+                  arrayOopDesc::base_offset_in_bytes(T_INT)),
+          rax);
+}
+
+void TemplateTable::lastore() {
+  transition(ltos, vtos);
+  __ pop_i(rbx);
+  // rax,: low(value)
+  // rcx: array
+  // rdx: high(value)
+  index_check(rcx, rbx);  // prefer index in rbx,
+  // rbx,: index
+  __ movptr(Address(rcx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 0 * wordSize), rax);
+  NOT_LP64(__ movl(Address(rcx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 1 * wordSize), rdx));
+}
+
+
+void TemplateTable::fastore() {
+  transition(ftos, vtos);
+  __ pop_i(rbx);
+  // xmm0: value
+  // rbx:  index
+  // rdx:  array
+  index_check(rdx, rbx); // prefer index in rbx
+  LP64_ONLY(__ movflt(Address(rdx, rbx,
+                   Address::times_4,
+                   arrayOopDesc::base_offset_in_bytes(T_FLOAT)),
+           xmm0));
+  NOT_LP64(__ fstp_s(Address(rdx, rbx, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_FLOAT))));
+}
+
+void TemplateTable::dastore() {
+  transition(dtos, vtos);
+  __ pop_i(rbx);
+  // xmm0: value
+  // rbx:  index
+  // rdx:  array
+  index_check(rdx, rbx); // prefer index in rbx
+  LP64_ONLY(__ movdbl(Address(rdx, rbx,
+                   Address::times_8,
+                   arrayOopDesc::base_offset_in_bytes(T_DOUBLE)),
+           xmm0));
+  NOT_LP64(__ fstp_d(Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_DOUBLE))));
+}
+
+void TemplateTable::aastore() {
+  Label is_null, ok_is_subtype, done;
+  transition(vtos, vtos);
+  // stack: ..., array, index, value
+  __ movptr(rax, at_tos());    // value
+  __ movl(rcx, at_tos_p1()); // index
+  __ movptr(rdx, at_tos_p2()); // array
+
+  Address element_address(rdx, rcx,
+                          UseCompressedOops? Address::times_4 : Address::times_ptr,
+                          arrayOopDesc::base_offset_in_bytes(T_OBJECT));
+
+  index_check_without_pop(rdx, rcx);     // kills rbx
+  __ testptr(rax, rax);
+  __ jcc(Assembler::zero, is_null);
+
+  // Move subklass into rbx
+  __ load_klass(rbx, rax);
+  // Move superklass into rax
+  __ load_klass(rax, rdx);
+  __ movptr(rax, Address(rax,
+                         ObjArrayKlass::element_klass_offset()));
+  // Compress array + index*oopSize + 12 into a single register.  Frees rcx.
+  __ lea(rdx, element_address);
+
+  // Generate subtype check.  Blows rcx, rdi
+  // Superklass in rax.  Subklass in rbx.
+  __ gen_subtype_check(rbx, ok_is_subtype);
+
+  // Come here on failure
+  // object is at TOS
+  __ jump(ExternalAddress(Interpreter::_throw_ArrayStoreException_entry));
+
+  // Come here on success
+  __ bind(ok_is_subtype);
+
+  // Get the value we will store
+  __ movptr(rax, at_tos());
+  // Now store using the appropriate barrier
+  do_oop_store(_masm, Address(rdx, 0), rax, _bs->kind(), true);
+  __ jmp(done);
+
+  // Have a NULL in rax, rdx=array, ecx=index.  Store NULL at ary[idx]
+  __ bind(is_null);
+  __ profile_null_seen(rbx);
+
+  // Store a NULL
+  do_oop_store(_masm, element_address, noreg, _bs->kind(), true);
+
+  // Pop stack arguments
+  __ bind(done);
+  __ addptr(rsp, 3 * Interpreter::stackElementSize);
+}
+
+void TemplateTable::bastore() {
+  transition(itos, vtos);
+  __ pop_i(rbx);
+  // rax: value
+  // rbx: index
+  // rdx: array
+  index_check(rdx, rbx); // prefer index in rbx
+  __ movb(Address(rdx, rbx,
+                  Address::times_1,
+                  arrayOopDesc::base_offset_in_bytes(T_BYTE)),
+          rax);
+}
+
+void TemplateTable::castore() {
+  transition(itos, vtos);
+  __ pop_i(rbx);
+  // rax: value
+  // rbx: index
+  // rdx: array
+  index_check(rdx, rbx);  // prefer index in rbx
+  __ movw(Address(rdx, rbx,
+                  Address::times_2,
+                  arrayOopDesc::base_offset_in_bytes(T_CHAR)),
+          rax);
+}
+
+
+void TemplateTable::sastore() {
+  castore();
+}
+
+void TemplateTable::istore(int n) {
+  transition(itos, vtos);
+  __ movl(iaddress(n), rax);
+}
+
+void TemplateTable::lstore(int n) {
+  transition(ltos, vtos);
+  __ movptr(laddress(n), rax);
+  NOT_LP64(__ movptr(haddress(n), rdx));
+}
+
+void TemplateTable::fstore(int n) {
+  transition(ftos, vtos);
+  LP64_ONLY(__ movflt(faddress(n), xmm0));
+  NOT_LP64(__ fstp_s(faddress(n)));
+}
+
+void TemplateTable::dstore(int n) {
+  transition(dtos, vtos);
+  LP64_ONLY(__ movdbl(daddress(n), xmm0));
+  NOT_LP64(__ fstp_d(daddress(n)));
+}
+
+
+void TemplateTable::astore(int n) {
+  transition(vtos, vtos);
+  __ pop_ptr(rax);
+  __ movptr(aaddress(n), rax);
+}
+
+void TemplateTable::pop() {
+  transition(vtos, vtos);
+  __ addptr(rsp, Interpreter::stackElementSize);
+}
+
+void TemplateTable::pop2() {
+  transition(vtos, vtos);
+  __ addptr(rsp, 2 * Interpreter::stackElementSize);
+}
+
+
+void TemplateTable::dup() {
+  transition(vtos, vtos);
+  __ load_ptr(0, rax);
+  __ push_ptr(rax);
+  // stack: ..., a, a
+}
+
+void TemplateTable::dup_x1() {
+  transition(vtos, vtos);
+  // stack: ..., a, b
+  __ load_ptr( 0, rax);  // load b
+  __ load_ptr( 1, rcx);  // load a
+  __ store_ptr(1, rax);  // store b
+  __ store_ptr(0, rcx);  // store a
+  __ push_ptr(rax);      // push b
+  // stack: ..., b, a, b
+}
+
+void TemplateTable::dup_x2() {
+  transition(vtos, vtos);
+  // stack: ..., a, b, c
+  __ load_ptr( 0, rax);  // load c
+  __ load_ptr( 2, rcx);  // load a
+  __ store_ptr(2, rax);  // store c in a
+  __ push_ptr(rax);      // push c
+  // stack: ..., c, b, c, c
+  __ load_ptr( 2, rax);  // load b
+  __ store_ptr(2, rcx);  // store a in b
+  // stack: ..., c, a, c, c
+  __ store_ptr(1, rax);  // store b in c
+  // stack: ..., c, a, b, c
+}
+
+void TemplateTable::dup2() {
+  transition(vtos, vtos);
+  // stack: ..., a, b
+  __ load_ptr(1, rax);  // load a
+  __ push_ptr(rax);     // push a
+  __ load_ptr(1, rax);  // load b
+  __ push_ptr(rax);     // push b
+  // stack: ..., a, b, a, b
+}
+
+
+void TemplateTable::dup2_x1() {
+  transition(vtos, vtos);
+  // stack: ..., a, b, c
+  __ load_ptr( 0, rcx);  // load c
+  __ load_ptr( 1, rax);  // load b
+  __ push_ptr(rax);      // push b
+  __ push_ptr(rcx);      // push c
+  // stack: ..., a, b, c, b, c
+  __ store_ptr(3, rcx);  // store c in b
+  // stack: ..., a, c, c, b, c
+  __ load_ptr( 4, rcx);  // load a
+  __ store_ptr(2, rcx);  // store a in 2nd c
+  // stack: ..., a, c, a, b, c
+  __ store_ptr(4, rax);  // store b in a
+  // stack: ..., b, c, a, b, c
+}
+
+void TemplateTable::dup2_x2() {
+  transition(vtos, vtos);
+  // stack: ..., a, b, c, d
+  __ load_ptr( 0, rcx);  // load d
+  __ load_ptr( 1, rax);  // load c
+  __ push_ptr(rax);      // push c
+  __ push_ptr(rcx);      // push d
+  // stack: ..., a, b, c, d, c, d
+  __ load_ptr( 4, rax);  // load b
+  __ store_ptr(2, rax);  // store b in d
+  __ store_ptr(4, rcx);  // store d in b
+  // stack: ..., a, d, c, b, c, d
+  __ load_ptr( 5, rcx);  // load a
+  __ load_ptr( 3, rax);  // load c
+  __ store_ptr(3, rcx);  // store a in c
+  __ store_ptr(5, rax);  // store c in a
+  // stack: ..., c, d, a, b, c, d
+}
+
+void TemplateTable::swap() {
+  transition(vtos, vtos);
+  // stack: ..., a, b
+  __ load_ptr( 1, rcx);  // load a
+  __ load_ptr( 0, rax);  // load b
+  __ store_ptr(0, rcx);  // store a in b
+  __ store_ptr(1, rax);  // store b in a
+  // stack: ..., b, a
+}
+
+void TemplateTable::iop2(Operation op) {
+  transition(itos, itos);
+  switch (op) {
+  case add  :                    __ pop_i(rdx); __ addl (rax, rdx); break;
+  case sub  : __ movl(rdx, rax); __ pop_i(rax); __ subl (rax, rdx); break;
+  case mul  :                    __ pop_i(rdx); __ imull(rax, rdx); break;
+  case _and :                    __ pop_i(rdx); __ andl (rax, rdx); break;
+  case _or  :                    __ pop_i(rdx); __ orl  (rax, rdx); break;
+  case _xor :                    __ pop_i(rdx); __ xorl (rax, rdx); break;
+  case shl  : __ movl(rcx, rax); __ pop_i(rax); __ shll (rax);      break;
+  case shr  : __ movl(rcx, rax); __ pop_i(rax); __ sarl (rax);      break;
+  case ushr : __ movl(rcx, rax); __ pop_i(rax); __ shrl (rax);      break;
+  default   : ShouldNotReachHere();
+  }
+}
+
+void TemplateTable::lop2(Operation op) {
+  transition(ltos, ltos);
+#ifdef _LP64
+  switch (op) {
+  case add  :                    __ pop_l(rdx); __ addptr(rax, rdx); break;
+  case sub  : __ mov(rdx, rax);  __ pop_l(rax); __ subptr(rax, rdx); break;
+  case _and :                    __ pop_l(rdx); __ andptr(rax, rdx); break;
+  case _or  :                    __ pop_l(rdx); __ orptr (rax, rdx); break;
+  case _xor :                    __ pop_l(rdx); __ xorptr(rax, rdx); break;
+  default   : ShouldNotReachHere();
+  }
+#else
+  __ pop_l(rbx, rcx);
+  switch (op) {
+    case add  : __ addl(rax, rbx); __ adcl(rdx, rcx); break;
+    case sub  : __ subl(rbx, rax); __ sbbl(rcx, rdx);
+                __ mov (rax, rbx); __ mov (rdx, rcx); break;
+    case _and : __ andl(rax, rbx); __ andl(rdx, rcx); break;
+    case _or  : __ orl (rax, rbx); __ orl (rdx, rcx); break;
+    case _xor : __ xorl(rax, rbx); __ xorl(rdx, rcx); break;
+    default   : ShouldNotReachHere();
+  }
+#endif
+}
+
+void TemplateTable::idiv() {
+  transition(itos, itos);
+  __ movl(rcx, rax);
+  __ pop_i(rax);
+  // Note: could xor rax and ecx and compare with (-1 ^ min_int). If
+  //       they are not equal, one could do a normal division (no correction
+  //       needed), which may speed up this implementation for the common case.
+  //       (see also JVM spec., p.243 & p.271)
+  __ corrected_idivl(rcx);
+}
+
+void TemplateTable::irem() {
+  transition(itos, itos);
+  __ movl(rcx, rax);
+  __ pop_i(rax);
+  // Note: could xor rax and ecx and compare with (-1 ^ min_int). If
+  //       they are not equal, one could do a normal division (no correction
+  //       needed), which may speed up this implementation for the common case.
+  //       (see also JVM spec., p.243 & p.271)
+  __ corrected_idivl(rcx);
+  __ movl(rax, rdx);
+}
+
+void TemplateTable::lmul() {
+  transition(ltos, ltos);
+#ifdef _LP64
+  __ pop_l(rdx);
+  __ imulq(rax, rdx);
+#else
+  __ pop_l(rbx, rcx);
+  __ push(rcx); __ push(rbx);
+  __ push(rdx); __ push(rax);
+  __ lmul(2 * wordSize, 0);
+  __ addptr(rsp, 4 * wordSize);  // take off temporaries
+#endif
+}
+
+void TemplateTable::ldiv() {
+  transition(ltos, ltos);
+#ifdef _LP64
+  __ mov(rcx, rax);
+  __ pop_l(rax);
+  // generate explicit div0 check
+  __ testq(rcx, rcx);
+  __ jump_cc(Assembler::zero,
+             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
+  // Note: could xor rax and rcx and compare with (-1 ^ min_int). If
+  //       they are not equal, one could do a normal division (no correction
+  //       needed), which may speed up this implementation for the common case.
+  //       (see also JVM spec., p.243 & p.271)
+  __ corrected_idivq(rcx); // kills rbx
+#else
+  __ pop_l(rbx, rcx);
+  __ push(rcx); __ push(rbx);
+  __ push(rdx); __ push(rax);
+  // check if y = 0
+  __ orl(rax, rdx);
+  __ jump_cc(Assembler::zero,
+             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
+  __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::ldiv));
+  __ addptr(rsp, 4 * wordSize);  // take off temporaries
+#endif
+}
+
+void TemplateTable::lrem() {
+  transition(ltos, ltos);
+#ifdef _LP64
+  __ mov(rcx, rax);
+  __ pop_l(rax);
+  __ testq(rcx, rcx);
+  __ jump_cc(Assembler::zero,
+             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
+  // Note: could xor rax and rcx and compare with (-1 ^ min_int). If
+  //       they are not equal, one could do a normal division (no correction
+  //       needed), which may speed up this implementation for the common case.
+  //       (see also JVM spec., p.243 & p.271)
+  __ corrected_idivq(rcx); // kills rbx
+  __ mov(rax, rdx);
+#else
+  __ pop_l(rbx, rcx);
+  __ push(rcx); __ push(rbx);
+  __ push(rdx); __ push(rax);
+  // check if y = 0
+  __ orl(rax, rdx);
+  __ jump_cc(Assembler::zero,
+             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
+  __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::lrem));
+  __ addptr(rsp, 4 * wordSize);
+#endif
+}
+
+void TemplateTable::lshl() {
+  transition(itos, ltos);
+  __ movl(rcx, rax);                             // get shift count
+  #ifdef _LP64
+  __ pop_l(rax);                                 // get shift value
+  __ shlq(rax);
+#else
+  __ pop_l(rax, rdx);                            // get shift value
+  __ lshl(rdx, rax);
+#endif
+}
+
+void TemplateTable::lshr() {
+#ifdef _LP64
+  transition(itos, ltos);
+  __ movl(rcx, rax);                             // get shift count
+  __ pop_l(rax);                                 // get shift value
+  __ sarq(rax);
+#else
+  transition(itos, ltos);
+  __ mov(rcx, rax);                              // get shift count
+  __ pop_l(rax, rdx);                            // get shift value
+  __ lshr(rdx, rax, true);
+#endif
+}
+
+void TemplateTable::lushr() {
+  transition(itos, ltos);
+#ifdef _LP64
+  __ movl(rcx, rax);                             // get shift count
+  __ pop_l(rax);                                 // get shift value
+  __ shrq(rax);
+#else
+  __ mov(rcx, rax);                              // get shift count
+  __ pop_l(rax, rdx);                            // get shift value
+  __ lshr(rdx, rax);
+#endif
+}
+
+void TemplateTable::fop2(Operation op) {
+  transition(ftos, ftos);
+#ifdef _LP64
+  switch (op) {
+  case add:
+    __ addss(xmm0, at_rsp());
+    __ addptr(rsp, Interpreter::stackElementSize);
+    break;
+  case sub:
+    __ movflt(xmm1, xmm0);
+    __ pop_f(xmm0);
+    __ subss(xmm0, xmm1);
+    break;
+  case mul:
+    __ mulss(xmm0, at_rsp());
+    __ addptr(rsp, Interpreter::stackElementSize);
+    break;
+  case div:
+    __ movflt(xmm1, xmm0);
+    __ pop_f(xmm0);
+    __ divss(xmm0, xmm1);
+    break;
+  case rem:
+    __ movflt(xmm1, xmm0);
+    __ pop_f(xmm0);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::frem), 2);
+    break;
+  default:
+    ShouldNotReachHere();
+    break;
+  }
+#else
+  switch (op) {
+    case add: __ fadd_s (at_rsp());                break;
+    case sub: __ fsubr_s(at_rsp());                break;
+    case mul: __ fmul_s (at_rsp());                break;
+    case div: __ fdivr_s(at_rsp());                break;
+    case rem: __ fld_s  (at_rsp()); __ fremr(rax); break;
+    default : ShouldNotReachHere();
+  }
+  __ f2ieee();
+  __ pop(rax);  // pop float thing off
+#endif
+}
+
+void TemplateTable::dop2(Operation op) {
+  transition(dtos, dtos);
+#ifdef _LP64
+  switch (op) {
+  case add:
+    __ addsd(xmm0, at_rsp());
+    __ addptr(rsp, 2 * Interpreter::stackElementSize);
+    break;
+  case sub:
+    __ movdbl(xmm1, xmm0);
+    __ pop_d(xmm0);
+    __ subsd(xmm0, xmm1);
+    break;
+  case mul:
+    __ mulsd(xmm0, at_rsp());
+    __ addptr(rsp, 2 * Interpreter::stackElementSize);
+    break;
+  case div:
+    __ movdbl(xmm1, xmm0);
+    __ pop_d(xmm0);
+    __ divsd(xmm0, xmm1);
+    break;
+  case rem:
+    __ movdbl(xmm1, xmm0);
+    __ pop_d(xmm0);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::drem), 2);
+    break;
+  default:
+    ShouldNotReachHere();
+    break;
+  }
+#else
+  switch (op) {
+    case add: __ fadd_d (at_rsp());                break;
+    case sub: __ fsubr_d(at_rsp());                break;
+    case mul: {
+      Label L_strict;
+      Label L_join;
+      const Address access_flags      (rcx, Method::access_flags_offset());
+      __ get_method(rcx);
+      __ movl(rcx, access_flags);
+      __ testl(rcx, JVM_ACC_STRICT);
+      __ jccb(Assembler::notZero, L_strict);
+      __ fmul_d (at_rsp());
+      __ jmpb(L_join);
+      __ bind(L_strict);
+      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias1()));
+      __ fmulp();
+      __ fmul_d (at_rsp());
+      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias2()));
+      __ fmulp();
+      __ bind(L_join);
+      break;
+    }
+    case div: {
+      Label L_strict;
+      Label L_join;
+      const Address access_flags      (rcx, Method::access_flags_offset());
+      __ get_method(rcx);
+      __ movl(rcx, access_flags);
+      __ testl(rcx, JVM_ACC_STRICT);
+      __ jccb(Assembler::notZero, L_strict);
+      __ fdivr_d(at_rsp());
+      __ jmp(L_join);
+      __ bind(L_strict);
+      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias1()));
+      __ fmul_d (at_rsp());
+      __ fdivrp();
+      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias2()));
+      __ fmulp();
+      __ bind(L_join);
+      break;
+    }
+    case rem: __ fld_d  (at_rsp()); __ fremr(rax); break;
+    default : ShouldNotReachHere();
+  }
+  __ d2ieee();
+  // Pop double precision number from rsp.
+  __ pop(rax);
+  __ pop(rdx);
+#endif
+}
+
+void TemplateTable::ineg() {
+  transition(itos, itos);
+  __ negl(rax);
+}
+
+void TemplateTable::lneg() {
+  transition(ltos, ltos);
+  LP64_ONLY(__ negq(rax));
+  NOT_LP64(__ lneg(rdx, rax));
+}
+
+#ifdef _LP64
+// Note: 'double' and 'long long' have 32-bits alignment on x86.
+static jlong* double_quadword(jlong *adr, jlong lo, jlong hi) {
+  // Use the expression (adr)&(~0xF) to provide 128-bits aligned address
+  // of 128-bits operands for SSE instructions.
+  jlong *operand = (jlong*)(((intptr_t)adr)&((intptr_t)(~0xF)));
+  // Store the value to a 128-bits operand.
+  operand[0] = lo;
+  operand[1] = hi;
+  return operand;
+}
+
+// Buffer for 128-bits masks used by SSE instructions.
+static jlong float_signflip_pool[2*2];
+static jlong double_signflip_pool[2*2];
+#endif
+
+void TemplateTable::fneg() {
+  transition(ftos, ftos);
+#ifdef _LP64
+  static jlong *float_signflip  = double_quadword(&float_signflip_pool[1], 0x8000000080000000, 0x8000000080000000);
+  __ xorps(xmm0, ExternalAddress((address) float_signflip));
+#else
+  __ fchs();
+#endif
+}
+
+void TemplateTable::dneg() {
+  transition(dtos, dtos);
+#ifdef _LP64
+  static jlong *double_signflip  = double_quadword(&double_signflip_pool[1], 0x8000000000000000, 0x8000000000000000);
+  __ xorpd(xmm0, ExternalAddress((address) double_signflip));
+#else
+  __ fchs();
+#endif
+}
+
+void TemplateTable::iinc() {
+  transition(vtos, vtos);
+  __ load_signed_byte(rdx, at_bcp(2)); // get constant
+  locals_index(rbx);
+  __ addl(iaddress(rbx), rdx);
+}
+
+void TemplateTable::wide_iinc() {
+  transition(vtos, vtos);
+  __ movl(rdx, at_bcp(4)); // get constant
+  locals_index_wide(rbx);
+  __ bswapl(rdx); // swap bytes & sign-extend constant
+  __ sarl(rdx, 16);
+  __ addl(iaddress(rbx), rdx);
+  // Note: should probably use only one movl to get both
+  //       the index and the constant -> fix this
+}
+
+void TemplateTable::convert() {
+#ifdef _LP64
+  // Checking
+#ifdef ASSERT
+  {
+    TosState tos_in  = ilgl;
+    TosState tos_out = ilgl;
+    switch (bytecode()) {
+    case Bytecodes::_i2l: // fall through
+    case Bytecodes::_i2f: // fall through
+    case Bytecodes::_i2d: // fall through
+    case Bytecodes::_i2b: // fall through
+    case Bytecodes::_i2c: // fall through
+    case Bytecodes::_i2s: tos_in = itos; break;
+    case Bytecodes::_l2i: // fall through
+    case Bytecodes::_l2f: // fall through
+    case Bytecodes::_l2d: tos_in = ltos; break;
+    case Bytecodes::_f2i: // fall through
+    case Bytecodes::_f2l: // fall through
+    case Bytecodes::_f2d: tos_in = ftos; break;
+    case Bytecodes::_d2i: // fall through
+    case Bytecodes::_d2l: // fall through
+    case Bytecodes::_d2f: tos_in = dtos; break;
+    default             : ShouldNotReachHere();
+    }
+    switch (bytecode()) {
+    case Bytecodes::_l2i: // fall through
+    case Bytecodes::_f2i: // fall through
+    case Bytecodes::_d2i: // fall through
+    case Bytecodes::_i2b: // fall through
+    case Bytecodes::_i2c: // fall through
+    case Bytecodes::_i2s: tos_out = itos; break;
+    case Bytecodes::_i2l: // fall through
+    case Bytecodes::_f2l: // fall through
+    case Bytecodes::_d2l: tos_out = ltos; break;
+    case Bytecodes::_i2f: // fall through
+    case Bytecodes::_l2f: // fall through
+    case Bytecodes::_d2f: tos_out = ftos; break;
+    case Bytecodes::_i2d: // fall through
+    case Bytecodes::_l2d: // fall through
+    case Bytecodes::_f2d: tos_out = dtos; break;
+    default             : ShouldNotReachHere();
+    }
+    transition(tos_in, tos_out);
+  }
+#endif // ASSERT
+
+  static const int64_t is_nan = 0x8000000000000000L;
+
+  // Conversion
+  switch (bytecode()) {
+  case Bytecodes::_i2l:
+    __ movslq(rax, rax);
+    break;
+  case Bytecodes::_i2f:
+    __ cvtsi2ssl(xmm0, rax);
+    break;
+  case Bytecodes::_i2d:
+    __ cvtsi2sdl(xmm0, rax);
+    break;
+  case Bytecodes::_i2b:
+    __ movsbl(rax, rax);
+    break;
+  case Bytecodes::_i2c:
+    __ movzwl(rax, rax);
+    break;
+  case Bytecodes::_i2s:
+    __ movswl(rax, rax);
+    break;
+  case Bytecodes::_l2i:
+    __ movl(rax, rax);
+    break;
+  case Bytecodes::_l2f:
+    __ cvtsi2ssq(xmm0, rax);
+    break;
+  case Bytecodes::_l2d:
+    __ cvtsi2sdq(xmm0, rax);
+    break;
+  case Bytecodes::_f2i:
+  {
+    Label L;
+    __ cvttss2sil(rax, xmm0);
+    __ cmpl(rax, 0x80000000); // NaN or overflow/underflow?
+    __ jcc(Assembler::notEqual, L);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2i), 1);
+    __ bind(L);
+  }
+    break;
+  case Bytecodes::_f2l:
+  {
+    Label L;
+    __ cvttss2siq(rax, xmm0);
+    // NaN or overflow/underflow?
+    __ cmp64(rax, ExternalAddress((address) &is_nan));
+    __ jcc(Assembler::notEqual, L);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2l), 1);
+    __ bind(L);
+  }
+    break;
+  case Bytecodes::_f2d:
+    __ cvtss2sd(xmm0, xmm0);
+    break;
+  case Bytecodes::_d2i:
+  {
+    Label L;
+    __ cvttsd2sil(rax, xmm0);
+    __ cmpl(rax, 0x80000000); // NaN or overflow/underflow?
+    __ jcc(Assembler::notEqual, L);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2i), 1);
+    __ bind(L);
+  }
+    break;
+  case Bytecodes::_d2l:
+  {
+    Label L;
+    __ cvttsd2siq(rax, xmm0);
+    // NaN or overflow/underflow?
+    __ cmp64(rax, ExternalAddress((address) &is_nan));
+    __ jcc(Assembler::notEqual, L);
+    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2l), 1);
+    __ bind(L);
+  }
+    break;
+  case Bytecodes::_d2f:
+    __ cvtsd2ss(xmm0, xmm0);
+    break;
+  default:
+    ShouldNotReachHere();
+  }
+#else
+  // Checking
+#ifdef ASSERT
+  { TosState tos_in  = ilgl;
+    TosState tos_out = ilgl;
+    switch (bytecode()) {
+      case Bytecodes::_i2l: // fall through
+      case Bytecodes::_i2f: // fall through
+      case Bytecodes::_i2d: // fall through
+      case Bytecodes::_i2b: // fall through
+      case Bytecodes::_i2c: // fall through
+      case Bytecodes::_i2s: tos_in = itos; break;
+      case Bytecodes::_l2i: // fall through
+      case Bytecodes::_l2f: // fall through
+      case Bytecodes::_l2d: tos_in = ltos; break;
+      case Bytecodes::_f2i: // fall through
+      case Bytecodes::_f2l: // fall through
+      case Bytecodes::_f2d: tos_in = ftos; break;
+      case Bytecodes::_d2i: // fall through
+      case Bytecodes::_d2l: // fall through
+      case Bytecodes::_d2f: tos_in = dtos; break;
+      default             : ShouldNotReachHere();
+    }
+    switch (bytecode()) {
+      case Bytecodes::_l2i: // fall through
+      case Bytecodes::_f2i: // fall through
+      case Bytecodes::_d2i: // fall through
+      case Bytecodes::_i2b: // fall through
+      case Bytecodes::_i2c: // fall through
+      case Bytecodes::_i2s: tos_out = itos; break;
+      case Bytecodes::_i2l: // fall through
+      case Bytecodes::_f2l: // fall through
+      case Bytecodes::_d2l: tos_out = ltos; break;
+      case Bytecodes::_i2f: // fall through
+      case Bytecodes::_l2f: // fall through
+      case Bytecodes::_d2f: tos_out = ftos; break;
+      case Bytecodes::_i2d: // fall through
+      case Bytecodes::_l2d: // fall through
+      case Bytecodes::_f2d: tos_out = dtos; break;
+      default             : ShouldNotReachHere();
+    }
+    transition(tos_in, tos_out);
+  }
+#endif // ASSERT
+
+  // Conversion
+  // (Note: use push(rcx)/pop(rcx) for 1/2-word stack-ptr manipulation)
+  switch (bytecode()) {
+    case Bytecodes::_i2l:
+      __ extend_sign(rdx, rax);
+      break;
+    case Bytecodes::_i2f:
+      __ push(rax);          // store int on tos
+      __ fild_s(at_rsp());   // load int to ST0
+      __ f2ieee();           // truncate to float size
+      __ pop(rcx);           // adjust rsp
+      break;
+    case Bytecodes::_i2d:
+      __ push(rax);          // add one slot for d2ieee()
+      __ push(rax);          // store int on tos
+      __ fild_s(at_rsp());   // load int to ST0
+      __ d2ieee();           // truncate to double size
+      __ pop(rcx);           // adjust rsp
+      __ pop(rcx);
+      break;
+    case Bytecodes::_i2b:
+      __ shll(rax, 24);      // truncate upper 24 bits
+      __ sarl(rax, 24);      // and sign-extend byte
+      LP64_ONLY(__ movsbl(rax, rax));
+      break;
+    case Bytecodes::_i2c:
+      __ andl(rax, 0xFFFF);  // truncate upper 16 bits
+      LP64_ONLY(__ movzwl(rax, rax));
+      break;
+    case Bytecodes::_i2s:
+      __ shll(rax, 16);      // truncate upper 16 bits
+      __ sarl(rax, 16);      // and sign-extend short
+      LP64_ONLY(__ movswl(rax, rax));
+      break;
+    case Bytecodes::_l2i:
+      /* nothing to do */
+      break;
+    case Bytecodes::_l2f:
+      __ push(rdx);          // store long on tos
+      __ push(rax);
+      __ fild_d(at_rsp());   // load long to ST0
+      __ f2ieee();           // truncate to float size
+      __ pop(rcx);           // adjust rsp
+      __ pop(rcx);
+      break;
+    case Bytecodes::_l2d:
+      __ push(rdx);          // store long on tos
+      __ push(rax);
+      __ fild_d(at_rsp());   // load long to ST0
+      __ d2ieee();           // truncate to double size
+      __ pop(rcx);           // adjust rsp
+      __ pop(rcx);
+      break;
+    case Bytecodes::_f2i:
+      __ push(rcx);          // reserve space for argument
+      __ fstp_s(at_rsp());   // pass float argument on stack
+      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2i), 1);
+      break;
+    case Bytecodes::_f2l:
+      __ push(rcx);          // reserve space for argument
+      __ fstp_s(at_rsp());   // pass float argument on stack
+      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2l), 1);
+      break;
+    case Bytecodes::_f2d:
+      /* nothing to do */
+      break;
+    case Bytecodes::_d2i:
+      __ push(rcx);          // reserve space for argument
+      __ push(rcx);
+      __ fstp_d(at_rsp());   // pass double argument on stack
+      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2i), 2);
+      break;
+    case Bytecodes::_d2l:
+      __ push(rcx);          // reserve space for argument
+      __ push(rcx);
+      __ fstp_d(at_rsp());   // pass double argument on stack
+      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2l), 2);
+      break;
+    case Bytecodes::_d2f:
+      __ push(rcx);          // reserve space for f2ieee()
+      __ f2ieee();           // truncate to float size
+      __ pop(rcx);           // adjust rsp
+      break;
+    default             :
+      ShouldNotReachHere();
+  }
+#endif
+}
+
+void TemplateTable::lcmp() {
+  transition(ltos, itos);
+#ifdef _LP64
+  Label done;
+  __ pop_l(rdx);
+  __ cmpq(rdx, rax);
+  __ movl(rax, -1);
+  __ jccb(Assembler::less, done);
+  __ setb(Assembler::notEqual, rax);
+  __ movzbl(rax, rax);
+  __ bind(done);
+#else
+
+  // y = rdx:rax
+  __ pop_l(rbx, rcx);             // get x = rcx:rbx
+  __ lcmp2int(rcx, rbx, rdx, rax);// rcx := cmp(x, y)
+  __ mov(rax, rcx);
+#endif
+}
+
+void TemplateTable::float_cmp(bool is_float, int unordered_result) {
+#ifdef _LP64
+  Label done;
+  if (is_float) {
+    // XXX get rid of pop here, use ... reg, mem32
+    __ pop_f(xmm1);
+    __ ucomiss(xmm1, xmm0);
+  } else {
+    // XXX get rid of pop here, use ... reg, mem64
+    __ pop_d(xmm1);
+    __ ucomisd(xmm1, xmm0);
+  }
+  if (unordered_result < 0) {
+    __ movl(rax, -1);
+    __ jccb(Assembler::parity, done);
+    __ jccb(Assembler::below, done);
+    __ setb(Assembler::notEqual, rdx);
+    __ movzbl(rax, rdx);
+  } else {
+    __ movl(rax, 1);
+    __ jccb(Assembler::parity, done);
+    __ jccb(Assembler::above, done);
+    __ movl(rax, 0);
+    __ jccb(Assembler::equal, done);
+    __ decrementl(rax);
+  }
+  __ bind(done);
+#else
+  if (is_float) {
+    __ fld_s(at_rsp());
+  } else {
+    __ fld_d(at_rsp());
+    __ pop(rdx);
+  }
+  __ pop(rcx);
+  __ fcmp2int(rax, unordered_result < 0);
+#endif
+}
+
+void TemplateTable::branch(bool is_jsr, bool is_wide) {
+  __ get_method(rcx); // rcx holds method
+  __ profile_taken_branch(rax, rbx); // rax holds updated MDP, rbx
+                                     // holds bumped taken count
+
+  const ByteSize be_offset = MethodCounters::backedge_counter_offset() +
+                             InvocationCounter::counter_offset();
+  const ByteSize inv_offset = MethodCounters::invocation_counter_offset() +
+                              InvocationCounter::counter_offset();
+
+  // Load up edx with the branch displacement
+  if (is_wide) {
+    __ movl(rdx, at_bcp(1));
+  } else {
+    __ load_signed_short(rdx, at_bcp(1));
+  }
+  __ bswapl(rdx);
+
+  if (!is_wide) {
+    __ sarl(rdx, 16);
+  }
+  LP64_ONLY(__ movl2ptr(rdx, rdx));
+
+  // Handle all the JSR stuff here, then exit.
+  // It's much shorter and cleaner than intermingling with the non-JSR
+  // normal-branch stuff occurring below.
+  if (is_jsr) {
+    // Pre-load the next target bytecode into rbx
+    __ load_unsigned_byte(rbx, Address(rbcp, rdx, Address::times_1, 0));
+
+    // compute return address as bci in rax
+    __ lea(rax, at_bcp((is_wide ? 5 : 3) -
+                        in_bytes(ConstMethod::codes_offset())));
+    __ subptr(rax, Address(rcx, Method::const_offset()));
+    // Adjust the bcp in r13 by the displacement in rdx
+    __ addptr(rbcp, rdx);
+    // jsr returns atos that is not an oop
+    __ push_i(rax);
+    __ dispatch_only(vtos);
+    return;
+  }
+
+  // Normal (non-jsr) branch handling
+
+  // Adjust the bcp in r13 by the displacement in rdx
+  __ addptr(rbcp, rdx);
+
+  assert(UseLoopCounter || !UseOnStackReplacement,
+         "on-stack-replacement requires loop counters");
+  Label backedge_counter_overflow;
+  Label profile_method;
+  Label dispatch;
+  if (UseLoopCounter) {
+    // increment backedge counter for backward branches
+    // rax: MDO
+    // rbx: MDO bumped taken-count
+    // rcx: method
+    // rdx: target offset
+    // r13: target bcp
+    // r14: locals pointer
+    __ testl(rdx, rdx);             // check if forward or backward branch
+    __ jcc(Assembler::positive, dispatch); // count only if backward branch
+
+    // check if MethodCounters exists
+    Label has_counters;
+    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
+    __ testptr(rax, rax);
+    __ jcc(Assembler::notZero, has_counters);
+    __ push(rdx);
+    __ push(rcx);
+    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::build_method_counters),
+               rcx);
+    __ pop(rcx);
+    __ pop(rdx);
+    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
+    __ jcc(Assembler::zero, dispatch);
+    __ bind(has_counters);
+
+    if (TieredCompilation) {
+      Label no_mdo;
+      int increment = InvocationCounter::count_increment;
+      if (ProfileInterpreter) {
+        // Are we profiling?
+        __ movptr(rbx, Address(rcx, in_bytes(Method::method_data_offset())));
+        __ testptr(rbx, rbx);
+        __ jccb(Assembler::zero, no_mdo);
+        // Increment the MDO backedge counter
+        const Address mdo_backedge_counter(rbx, in_bytes(MethodData::backedge_counter_offset()) +
+                                           in_bytes(InvocationCounter::counter_offset()));
+        const Address mask(rbx, in_bytes(MethodData::backedge_mask_offset()));
+        __ increment_mask_and_jump(mdo_backedge_counter, increment, mask,
+                                   rax, false, Assembler::zero, &backedge_counter_overflow);
+        __ jmp(dispatch);
+      }
+      __ bind(no_mdo);
+      // Increment backedge counter in MethodCounters*
+      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
+      const Address mask(rcx, in_bytes(MethodCounters::backedge_mask_offset()));
+      __ increment_mask_and_jump(Address(rcx, be_offset), increment, mask,
+                                 rax, false, Assembler::zero, &backedge_counter_overflow);
+    } else { // not TieredCompilation
+      // increment counter
+      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
+      __ movl(rax, Address(rcx, be_offset));        // load backedge counter
+      __ incrementl(rax, InvocationCounter::count_increment); // increment counter
+      __ movl(Address(rcx, be_offset), rax);        // store counter
+
+      __ movl(rax, Address(rcx, inv_offset));    // load invocation counter
+
+      __ andl(rax, InvocationCounter::count_mask_value); // and the status bits
+      __ addl(rax, Address(rcx, be_offset));        // add both counters
+
+      if (ProfileInterpreter) {
+        // Test to see if we should create a method data oop
+        __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_profile_limit_offset())));
+        __ jcc(Assembler::less, dispatch);
+
+        // if no method data exists, go to profile method
+        __ test_method_data_pointer(rax, profile_method);
+
+        if (UseOnStackReplacement) {
+          // check for overflow against rbx which is the MDO taken count
+          __ cmp32(rbx, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
+          __ jcc(Assembler::below, dispatch);
+
+          // When ProfileInterpreter is on, the backedge_count comes
+          // from the MethodData*, which value does not get reset on
+          // the call to frequency_counter_overflow().  To avoid
+          // excessive calls to the overflow routine while the method is
+          // being compiled, add a second test to make sure the overflow
+          // function is called only once every overflow_frequency.
+          const int overflow_frequency = 1024;
+          __ andl(rbx, overflow_frequency - 1);
+          __ jcc(Assembler::zero, backedge_counter_overflow);
+
+        }
+      } else {
+        if (UseOnStackReplacement) {
+          // check for overflow against rax, which is the sum of the
+          // counters
+          __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
+          __ jcc(Assembler::aboveEqual, backedge_counter_overflow);
+
+        }
+      }
+    }
+    __ bind(dispatch);
+  }
+
+  // Pre-load the next target bytecode into rbx
+  __ load_unsigned_byte(rbx, Address(rbcp, 0));
+
+  // continue with the bytecode @ target
+  // rax: return bci for jsr's, unused otherwise
+  // rbx: target bytecode
+  // r13: target bcp
+  __ dispatch_only(vtos);
+
+  if (UseLoopCounter) {
+    if (ProfileInterpreter) {
+      // Out-of-line code to allocate method data oop.
+      __ bind(profile_method);
+      __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::profile_method));
+      __ load_unsigned_byte(rbx, Address(rbcp, 0));  // restore target bytecode
+      __ set_method_data_pointer_for_bcp();
+      __ jmp(dispatch);
+    }
+
+    if (UseOnStackReplacement) {
+      // invocation counter overflow
+      __ bind(backedge_counter_overflow);
+      __ negptr(rdx);
+      __ addptr(rdx, rbcp); // branch bcp
+      // IcoResult frequency_counter_overflow([JavaThread*], address branch_bcp)
+      __ call_VM(noreg,
+                 CAST_FROM_FN_PTR(address,
+                                  InterpreterRuntime::frequency_counter_overflow),
+                 rdx);
+      __ load_unsigned_byte(rbx, Address(rbcp, 0));  // restore target bytecode
+
+      // rax: osr nmethod (osr ok) or NULL (osr not possible)
+      // rbx: target bytecode
+      // rdx: scratch
+      // r14: locals pointer
+      // r13: bcp
+      __ testptr(rax, rax);                        // test result
+      __ jcc(Assembler::zero, dispatch);         // no osr if null
+      // nmethod may have been invalidated (VM may block upon call_VM return)
+      __ cmpb(Address(rax, nmethod::state_offset()), nmethod::in_use);
+      __ jcc(Assembler::notEqual, dispatch);
+
+      // We have the address of an on stack replacement routine in rax
+      // We need to prepare to execute the OSR method. First we must
+      // migrate the locals and monitors off of the stack.
+
+      LP64_ONLY(__ mov(r13, rax));                             // save the nmethod
+      NOT_LP64(__ mov(rbx, rax));                             // save the nmethod
+      NOT_LP64(__ get_thread(rcx));
+
+      call_VM(noreg, CAST_FROM_FN_PTR(address, SharedRuntime::OSR_migration_begin));
+
+      // rax is OSR buffer, move it to expected parameter location
+      LP64_ONLY(__ mov(j_rarg0, rax));
+      NOT_LP64(__ mov(rcx, rax));
+      // We use j_rarg definitions here so that registers don't conflict as parameter
+      // registers change across platforms as we are in the midst of a calling
+      // sequence to the OSR nmethod and we don't want collision. These are NOT parameters.
+
+      const Register retaddr   = LP64_ONLY(j_rarg2) NOT_LP64(rdi);
+      const Register sender_sp = LP64_ONLY(j_rarg1) NOT_LP64(rdx);
+
+
+      // pop the interpreter frame
+      __ movptr(sender_sp, Address(rbp, frame::interpreter_frame_sender_sp_offset * wordSize)); // get sender sp
+      __ leave();                                // remove frame anchor
+      __ pop(retaddr);                           // get return address
+      __ mov(rsp, sender_sp);                   // set sp to sender sp
+      // Ensure compiled code always sees stack at proper alignment
+      __ andptr(rsp, -(StackAlignmentInBytes));
+
+      // unlike x86 we need no specialized return from compiled code
+      // to the interpreter or the call stub.
+
+      // push the return address
+      __ push(retaddr);
+
+      // and begin the OSR nmethod
+      LP64_ONLY(__ jmp(Address(r13, nmethod::osr_entry_point_offset())));
+      NOT_LP64(__ jmp(Address(rbx, nmethod::osr_entry_point_offset())));
+    }
+  }
+}
+
+void TemplateTable::if_0cmp(Condition cc) {
+  transition(itos, vtos);
+  // assume branch is more often taken than not (loops use backward branches)
+  Label not_taken;
+  __ testl(rax, rax);
+  __ jcc(j_not(cc), not_taken);
+  branch(false, false);
+  __ bind(not_taken);
+  __ profile_not_taken_branch(rax);
+}
+
+void TemplateTable::if_icmp(Condition cc) {
+  transition(itos, vtos);
+  // assume branch is more often taken than not (loops use backward branches)
+  Label not_taken;
+  __ pop_i(rdx);
+  __ cmpl(rdx, rax);
+  __ jcc(j_not(cc), not_taken);
+  branch(false, false);
+  __ bind(not_taken);
+  __ profile_not_taken_branch(rax);
+}
+
+void TemplateTable::if_nullcmp(Condition cc) {
+  transition(atos, vtos);
+  // assume branch is more often taken than not (loops use backward branches)
+  Label not_taken;
+  __ testptr(rax, rax);
+  __ jcc(j_not(cc), not_taken);
+  branch(false, false);
+  __ bind(not_taken);
+  __ profile_not_taken_branch(rax);
+}
+
+void TemplateTable::if_acmp(Condition cc) {
+  transition(atos, vtos);
+  // assume branch is more often taken than not (loops use backward branches)
+  Label not_taken;
+  __ pop_ptr(rdx);
+  __ cmpptr(rdx, rax);
+  __ jcc(j_not(cc), not_taken);
+  branch(false, false);
+  __ bind(not_taken);
+  __ profile_not_taken_branch(rax);
+}
+
+void TemplateTable::ret() {
+  transition(vtos, vtos);
+  locals_index(rbx);
+  LP64_ONLY(__ movslq(rbx, iaddress(rbx))); // get return bci, compute return bcp
+  NOT_LP64(__ movptr(rbx, iaddress(rbx)));
+  __ profile_ret(rbx, rcx);
+  __ get_method(rax);
+  __ movptr(rbcp, Address(rax, Method::const_offset()));
+  __ lea(rbcp, Address(rbcp, rbx, Address::times_1,
+                      ConstMethod::codes_offset()));
+  __ dispatch_next(vtos);
+}
+
+void TemplateTable::wide_ret() {
+  transition(vtos, vtos);
+  locals_index_wide(rbx);
+  __ movptr(rbx, aaddress(rbx)); // get return bci, compute return bcp
+  __ profile_ret(rbx, rcx);
+  __ get_method(rax);
+  __ movptr(rbcp, Address(rax, Method::const_offset()));
+  __ lea(rbcp, Address(rbcp, rbx, Address::times_1, ConstMethod::codes_offset()));
+  __ dispatch_next(vtos);
+}
+
+void TemplateTable::tableswitch() {
+  Label default_case, continue_execution;
+  transition(itos, vtos);
+
+  // align r13/rsi
+  __ lea(rbx, at_bcp(BytesPerInt));
+  __ andptr(rbx, -BytesPerInt);
+  // load lo & hi
+  __ movl(rcx, Address(rbx, BytesPerInt));
+  __ movl(rdx, Address(rbx, 2 * BytesPerInt));
+  __ bswapl(rcx);
+  __ bswapl(rdx);
+  // check against lo & hi
+  __ cmpl(rax, rcx);
+  __ jcc(Assembler::less, default_case);
+  __ cmpl(rax, rdx);
+  __ jcc(Assembler::greater, default_case);
+  // lookup dispatch offset
+  __ subl(rax, rcx);
+  __ movl(rdx, Address(rbx, rax, Address::times_4, 3 * BytesPerInt));
+  __ profile_switch_case(rax, rbx, rcx);
+  // continue execution
+  __ bind(continue_execution);
+  __ bswapl(rdx);
+  LP64_ONLY(__ movl2ptr(rdx, rdx));
+  __ load_unsigned_byte(rbx, Address(rbcp, rdx, Address::times_1));
+  __ addptr(rbcp, rdx);
+  __ dispatch_only(vtos);
+  // handle default
+  __ bind(default_case);
+  __ profile_switch_default(rax);
+  __ movl(rdx, Address(rbx, 0));
+  __ jmp(continue_execution);
+}
+
+void TemplateTable::lookupswitch() {
+  transition(itos, itos);
+  __ stop("lookupswitch bytecode should have been rewritten");
+}
+
+void TemplateTable::fast_linearswitch() {
+  transition(itos, vtos);
+  Label loop_entry, loop, found, continue_execution;
+  // bswap rax so we can avoid bswapping the table entries
+  __ bswapl(rax);
+  // align r13
+  __ lea(rbx, at_bcp(BytesPerInt)); // btw: should be able to get rid of
+                                    // this instruction (change offsets
+                                    // below)
+  __ andptr(rbx, -BytesPerInt);
+  // set counter
+  __ movl(rcx, Address(rbx, BytesPerInt));
+  __ bswapl(rcx);
+  __ jmpb(loop_entry);
+  // table search
+  __ bind(loop);
+  __ cmpl(rax, Address(rbx, rcx, Address::times_8, 2 * BytesPerInt));
+  __ jcc(Assembler::equal, found);
+  __ bind(loop_entry);
+  __ decrementl(rcx);
+  __ jcc(Assembler::greaterEqual, loop);
+  // default case
+  __ profile_switch_default(rax);
+  __ movl(rdx, Address(rbx, 0));
+  __ jmp(continue_execution);
+  // entry found -> get offset
+  __ bind(found);
+  __ movl(rdx, Address(rbx, rcx, Address::times_8, 3 * BytesPerInt));
+  __ profile_switch_case(rcx, rax, rbx);
+  // continue execution
+  __ bind(continue_execution);
+  __ bswapl(rdx);
+  __ movl2ptr(rdx, rdx);
+  __ load_unsigned_byte(rbx, Address(rbcp, rdx, Address::times_1));
+  __ addptr(rbcp, rdx);
+  __ dispatch_only(vtos);
+}
+
+void TemplateTable::fast_binaryswitch() {
+  transition(itos, vtos);
+  // Implementation using the following core algorithm:
+  //
+  // int binary_search(int key, LookupswitchPair* array, int n) {
+  //   // Binary search according to "Methodik des Programmierens" by
+  //   // Edsger W. Dijkstra and W.H.J. Feijen, Addison Wesley Germany 1985.
+  //   int i = 0;
+  //   int j = n;
+  //   while (i+1 < j) {
+  //     // invariant P: 0 <= i < j <= n and (a[i] <= key < a[j] or Q)
+  //     // with      Q: for all i: 0 <= i < n: key < a[i]
+  //     // where a stands for the array and assuming that the (inexisting)
+  //     // element a[n] is infinitely big.
+  //     int h = (i + j) >> 1;
+  //     // i < h < j
+  //     if (key < array[h].fast_match()) {
+  //       j = h;
+  //     } else {
+  //       i = h;
+  //     }
+  //   }
+  //   // R: a[i] <= key < a[i+1] or Q
+  //   // (i.e., if key is within array, i is the correct index)
+  //   return i;
+  // }
+
+  // Register allocation
+  const Register key   = rax; // already set (tosca)
+  const Register array = rbx;
+  const Register i     = rcx;
+  const Register j     = rdx;
+  const Register h     = rdi;
+  const Register temp  = rsi;
+
+  // Find array start
+  NOT_LP64(__ save_bcp());
+
+  __ lea(array, at_bcp(3 * BytesPerInt)); // btw: should be able to
+                                          // get rid of this
+                                          // instruction (change
+                                          // offsets below)
+  __ andptr(array, -BytesPerInt);
+
+  // Initialize i & j
+  __ xorl(i, i);                            // i = 0;
+  __ movl(j, Address(array, -BytesPerInt)); // j = length(array);
+
+  // Convert j into native byteordering
+  __ bswapl(j);
+
+  // And start
+  Label entry;
+  __ jmp(entry);
+
+  // binary search loop
+  {
+    Label loop;
+    __ bind(loop);
+    // int h = (i + j) >> 1;
+    __ leal(h, Address(i, j, Address::times_1)); // h = i + j;
+    __ sarl(h, 1);                               // h = (i + j) >> 1;
+    // if (key < array[h].fast_match()) {
+    //   j = h;
+    // } else {
+    //   i = h;
+    // }
+    // Convert array[h].match to native byte-ordering before compare
+    __ movl(temp, Address(array, h, Address::times_8));
+    __ bswapl(temp);
+    __ cmpl(key, temp);
+    // j = h if (key <  array[h].fast_match())
+    __ cmov32(Assembler::less, j, h);
+    // i = h if (key >= array[h].fast_match())
+    __ cmov32(Assembler::greaterEqual, i, h);
+    // while (i+1 < j)
+    __ bind(entry);
+    __ leal(h, Address(i, 1)); // i+1
+    __ cmpl(h, j);             // i+1 < j
+    __ jcc(Assembler::less, loop);
+  }
+
+  // end of binary search, result index is i (must check again!)
+  Label default_case;
+  // Convert array[i].match to native byte-ordering before compare
+  __ movl(temp, Address(array, i, Address::times_8));
+  __ bswapl(temp);
+  __ cmpl(key, temp);
+  __ jcc(Assembler::notEqual, default_case);
+
+  // entry found -> j = offset
+  __ movl(j , Address(array, i, Address::times_8, BytesPerInt));
+  __ profile_switch_case(i, key, array);
+  __ bswapl(j);
+  LP64_ONLY(__ movslq(j, j));
+
+  NOT_LP64(__ restore_bcp());
+  NOT_LP64(__ restore_locals());                           // restore rdi
+
+  __ load_unsigned_byte(rbx, Address(rbcp, j, Address::times_1));
+  __ addptr(rbcp, j);
+  __ dispatch_only(vtos);
+
+  // default case -> j = default offset
+  __ bind(default_case);
+  __ profile_switch_default(i);
+  __ movl(j, Address(array, -2 * BytesPerInt));
+  __ bswapl(j);
+  LP64_ONLY(__ movslq(j, j));
+
+  NOT_LP64(__ restore_bcp());
+  NOT_LP64(__ restore_locals());
+
+  __ load_unsigned_byte(rbx, Address(rbcp, j, Address::times_1));
+  __ addptr(rbcp, j);
+  __ dispatch_only(vtos);
+}
+
+void TemplateTable::_return(TosState state) {
+  transition(state, state);
+
+  Register robj = LP64_ONLY(c_rarg1) NOT_LP64(rax);
+
+  assert(_desc->calls_vm(),
+         "inconsistent calls_vm information"); // call in remove_activation
+
+  if (_desc->bytecode() == Bytecodes::_return_register_finalizer) {
+    assert(state == vtos, "only valid state");
+    __ movptr(robj, aaddress(0));
+    __ load_klass(rdi, robj);
+    __ movl(rdi, Address(rdi, Klass::access_flags_offset()));
+    __ testl(rdi, JVM_ACC_HAS_FINALIZER);
+    Label skip_register_finalizer;
+    __ jcc(Assembler::zero, skip_register_finalizer);
+
+    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::register_finalizer), robj);
+
+    __ bind(skip_register_finalizer);
+  }
+
+  __ remove_activation(state, rbcp);
+  __ jmp(rbcp);
+}
+
+// ----------------------------------------------------------------------------
+// Volatile variables demand their effects be made known to all CPU's
+// in order.  Store buffers on most chips allow reads & writes to
+// reorder; the JMM's ReadAfterWrite.java test fails in -Xint mode
+// without some kind of memory barrier (i.e., it's not sufficient that
+// the interpreter does not reorder volatile references, the hardware
+// also must not reorder them).
+//
+// According to the new Java Memory Model (JMM):
+// (1) All volatiles are serialized wrt to each other.  ALSO reads &
+//     writes act as aquire & release, so:
+// (2) A read cannot let unrelated NON-volatile memory refs that
+//     happen after the read float up to before the read.  It's OK for
+//     non-volatile memory refs that happen before the volatile read to
+//     float down below it.
+// (3) Similar a volatile write cannot let unrelated NON-volatile
+//     memory refs that happen BEFORE the write float down to after the
+//     write.  It's OK for non-volatile memory refs that happen after the
+//     volatile write to float up before it.
+//
+// We only put in barriers around volatile refs (they are expensive),
+// not _between_ memory refs (that would require us to track the
+// flavor of the previous memory refs).  Requirements (2) and (3)
+// require some barriers before volatile stores and after volatile
+// loads.  These nearly cover requirement (1) but miss the
+// volatile-store-volatile-load case.  This final case is placed after
+// volatile-stores although it could just as well go before
+// volatile-loads.
+
+void TemplateTable::volatile_barrier(Assembler::Membar_mask_bits order_constraint ) {
+  // Helper function to insert a is-volatile test and memory barrier
+  if(!os::is_MP()) return;    // Not needed on single CPU
+  __ membar(order_constraint);
+}
+
+void TemplateTable::resolve_cache_and_index(int byte_no,
+                                            Register Rcache,
+                                            Register index,
+                                            size_t index_size) {
+  const Register temp = rbx;
+  assert_different_registers(Rcache, index, temp);
+
+  Label resolved;
+    assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
+    __ get_cache_and_index_and_bytecode_at_bcp(Rcache, index, temp, byte_no, 1, index_size);
+    __ cmpl(temp, (int) bytecode());  // have we resolved this bytecode?
+    __ jcc(Assembler::equal, resolved);
+
+  // resolve first time through
+  address entry;
+  switch (bytecode()) {
+    case Bytecodes::_getstatic      : // fall through
+    case Bytecodes::_putstatic      : // fall through
+    case Bytecodes::_getfield       : // fall through
+    case Bytecodes::_putfield       : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_get_put);        break;
+    case Bytecodes::_invokevirtual  : // fall through
+    case Bytecodes::_invokespecial  : // fall through
+    case Bytecodes::_invokestatic   : // fall through
+    case Bytecodes::_invokeinterface: entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invoke);         break;
+    case Bytecodes::_invokehandle   : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokehandle);   break;
+    case Bytecodes::_invokedynamic  : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokedynamic);  break;
+    default:
+      fatal(err_msg("unexpected bytecode: %s", Bytecodes::name(bytecode())));
+      break;
+  }
+  __ movl(temp, (int)bytecode());
+  __ call_VM(noreg, entry, temp);
+  // Update registers with resolved info
+  __ get_cache_and_index_at_bcp(Rcache, index, 1, index_size);
+  __ bind(resolved);
+}
+
+// The cache and index registers must be set before call
+void TemplateTable::load_field_cp_cache_entry(Register obj,
+                                              Register cache,
+                                              Register index,
+                                              Register off,
+                                              Register flags,
+                                              bool is_static = false) {
+  assert_different_registers(cache, index, flags, off);
+
+  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
+  // Field offset
+  __ movptr(off, Address(cache, index, Address::times_ptr,
+                         in_bytes(cp_base_offset +
+                                  ConstantPoolCacheEntry::f2_offset())));
+  // Flags
+  __ movl(flags, Address(cache, index, Address::times_ptr,
+                         in_bytes(cp_base_offset +
+                                  ConstantPoolCacheEntry::flags_offset())));
+
+  // klass overwrite register
+  if (is_static) {
+    __ movptr(obj, Address(cache, index, Address::times_ptr,
+                           in_bytes(cp_base_offset +
+                                    ConstantPoolCacheEntry::f1_offset())));
+    const int mirror_offset = in_bytes(Klass::java_mirror_offset());
+    __ movptr(obj, Address(obj, mirror_offset));
+  }
+}
+
+void TemplateTable::load_invoke_cp_cache_entry(int byte_no,
+                                               Register method,
+                                               Register itable_index,
+                                               Register flags,
+                                               bool is_invokevirtual,
+                                               bool is_invokevfinal, /*unused*/
+                                               bool is_invokedynamic) {
+  // setup registers
+  const Register cache = rcx;
+  const Register index = rdx;
+  assert_different_registers(method, flags);
+  assert_different_registers(method, cache, index);
+  assert_different_registers(itable_index, flags);
+  assert_different_registers(itable_index, cache, index);
+  // determine constant pool cache field offsets
+  assert(is_invokevirtual == (byte_no == f2_byte), "is_invokevirtual flag redundant");
+  const int method_offset = in_bytes(
+    ConstantPoolCache::base_offset() +
+      ((byte_no == f2_byte)
+       ? ConstantPoolCacheEntry::f2_offset()
+       : ConstantPoolCacheEntry::f1_offset()));
+  const int flags_offset = in_bytes(ConstantPoolCache::base_offset() +
+                                    ConstantPoolCacheEntry::flags_offset());
+  // access constant pool cache fields
+  const int index_offset = in_bytes(ConstantPoolCache::base_offset() +
+                                    ConstantPoolCacheEntry::f2_offset());
+
+  size_t index_size = (is_invokedynamic ? sizeof(u4) : sizeof(u2));
+  resolve_cache_and_index(byte_no, cache, index, index_size);
+    __ movptr(method, Address(cache, index, Address::times_ptr, method_offset));
+
+  if (itable_index != noreg) {
+    // pick up itable or appendix index from f2 also:
+    __ movptr(itable_index, Address(cache, index, Address::times_ptr, index_offset));
+  }
+  __ movl(flags, Address(cache, index, Address::times_ptr, flags_offset));
+}
+
+// The registers cache and index expected to be set before call.
+// Correct values of the cache and index registers are preserved.
+void TemplateTable::jvmti_post_field_access(Register cache,
+                                            Register index,
+                                            bool is_static,
+                                            bool has_tos) {
+  if (JvmtiExport::can_post_field_access()) {
+    // Check to see if a field access watch has been set before we take
+    // the time to call into the VM.
+    Label L1;
+    assert_different_registers(cache, index, rax);
+    __ mov32(rax, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
+    __ testl(rax,rax);
+    __ jcc(Assembler::zero, L1);
+
+    // cache entry pointer
+    __ addptr(cache, in_bytes(ConstantPoolCache::base_offset()));
+    __ shll(index, LogBytesPerWord);
+    __ addptr(cache, index);
+    if (is_static) {
+      __ xorptr(rax, rax);      // NULL object reference
+    } else {
+      __ pop(atos);         // Get the object
+      __ verify_oop(rax);
+      __ push(atos);        // Restore stack state
+    }
+    // rax,:   object pointer or NULL
+    // cache: cache entry pointer
+    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_access),
+               rax, cache);
+    __ get_cache_and_index_at_bcp(cache, index, 1);
+    __ bind(L1);
+  }
+}
+
+void TemplateTable::pop_and_check_object(Register r) {
+  __ pop_ptr(r);
+  __ null_check(r);  // for field access must check obj.
+  __ verify_oop(r);
+}
+
+void TemplateTable::getfield_or_static(int byte_no, bool is_static) {
+  transition(vtos, vtos);
+
+  const Register cache = rcx;
+  const Register index = rdx;
+  const Register obj   = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+  const Register off   = rbx;
+  const Register flags = rax;
+  const Register bc    = LP64_ONLY(c_rarg3) NOT_LP64(rcx); // uses same reg as obj, so don't mix them
+
+  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
+  jvmti_post_field_access(cache, index, is_static, false);
+  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
+
+  if (!is_static) pop_and_check_object(obj);
+
+  const Address field(obj, off, Address::times_1, 0*wordSize);
+  NOT_LP64(const Address hi(obj, off, Address::times_1, 1*wordSize));
+
+  Label Done, notByte, notInt, notShort, notChar, notLong, notFloat, notObj, notDouble;
+
+  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
+  // Make sure we don't need to mask edx after the above shift
+  assert(btos == 0, "change code, btos != 0");
+
+  __ andl(flags, ConstantPoolCacheEntry::tos_state_mask);
+
+  __ jcc(Assembler::notZero, notByte);
+  // btos
+  __ load_signed_byte(rax, field);
+  __ push(btos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_bgetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notByte);
+  __ cmpl(flags, atos);
+  __ jcc(Assembler::notEqual, notObj);
+  // atos
+  __ load_heap_oop(rax, field);
+  __ push(atos);
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_agetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notObj);
+  __ cmpl(flags, itos);
+  __ jcc(Assembler::notEqual, notInt);
+  // itos
+  __ movl(rax, field);
+  __ push(itos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_igetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notInt);
+  __ cmpl(flags, ctos);
+  __ jcc(Assembler::notEqual, notChar);
+  // ctos
+  __ load_unsigned_short(rax, field);
+  __ push(ctos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_cgetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notChar);
+  __ cmpl(flags, stos);
+  __ jcc(Assembler::notEqual, notShort);
+  // stos
+  __ load_signed_short(rax, field);
+  __ push(stos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_sgetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notShort);
+  __ cmpl(flags, ltos);
+  __ jcc(Assembler::notEqual, notLong);
+  // ltos
+
+#ifndef _LP64
+  // Generate code as if volatile.  There just aren't enough registers to
+  // save that information and this code is faster than the test.
+  __ fild_d(field);                // Must load atomically
+  __ subptr(rsp,2*wordSize);    // Make space for store
+  __ fistp_d(Address(rsp,0));
+  __ pop(rax);
+  __ pop(rdx);
+#else
+  __ movq(rax, field);
+#endif
+
+  __ push(ltos);
+  // Rewrite bytecode to be faster
+  LP64_ONLY(if (!is_static) patch_bytecode(Bytecodes::_fast_lgetfield, bc, rbx));
+  __ jmp(Done);
+
+  __ bind(notLong);
+  __ cmpl(flags, ftos);
+  __ jcc(Assembler::notEqual, notFloat);
+  // ftos
+
+  LP64_ONLY(__ movflt(xmm0, field));
+  NOT_LP64(__ fld_s(field));
+  __ push(ftos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_fgetfield, bc, rbx);
+  }
+  __ jmp(Done);
+
+  __ bind(notFloat);
+#ifdef ASSERT
+  __ cmpl(flags, dtos);
+  __ jcc(Assembler::notEqual, notDouble);
+#endif
+  // dtos
+  LP64_ONLY(__ movdbl(xmm0, field));
+  NOT_LP64(__ fld_d(field));
+  __ push(dtos);
+  // Rewrite bytecode to be faster
+  if (!is_static) {
+    patch_bytecode(Bytecodes::_fast_dgetfield, bc, rbx);
+  }
+#ifdef ASSERT
+  __ jmp(Done);
+
+
+  __ bind(notDouble);
+  __ stop("Bad state");
+#endif
+
+  __ bind(Done);
+  // [jk] not needed currently
+  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadLoad |
+  //                                              Assembler::LoadStore));
+}
+
+void TemplateTable::getfield(int byte_no) {
+  getfield_or_static(byte_no, false);
+}
+
+void TemplateTable::getstatic(int byte_no) {
+  getfield_or_static(byte_no, true);
+}
+
+
+// The registers cache and index expected to be set before call.
+// The function may destroy various registers, just not the cache and index registers.
+void TemplateTable::jvmti_post_field_mod(Register cache, Register index, bool is_static) {
+
+  const Register robj = LP64_ONLY(c_rarg2)   NOT_LP64(rax);
+  const Register RBX  = LP64_ONLY(c_rarg1)   NOT_LP64(rbx);
+  const Register RCX  = LP64_ONLY(c_rarg3)   NOT_LP64(rcx);
+  const Register RDX  = LP64_ONLY(rscratch1) NOT_LP64(rdx);
+
+  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
+
+  if (JvmtiExport::can_post_field_modification()) {
+    // Check to see if a field modification watch has been set before
+    // we take the time to call into the VM.
+    Label L1;
+    assert_different_registers(cache, index, rax);
+    __ mov32(rax, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
+    __ testl(rax, rax);
+    __ jcc(Assembler::zero, L1);
+
+    __ get_cache_and_index_at_bcp(robj, RDX, 1);
+
+
+    if (is_static) {
+      // Life is simple.  Null out the object pointer.
+      __ xorl(RBX, RBX);
+
+    } else {
+      // Life is harder. The stack holds the value on top, followed by
+      // the object.  We don't know the size of the value, though; it
+      // could be one or two words depending on its type. As a result,
+      // we must find the type to determine where the object is.
+#ifndef _LP64
+      Label two_word, valsize_known;
+#endif
+      __ movl(RCX, Address(robj, RDX,
+                           Address::times_ptr,
+                           in_bytes(cp_base_offset +
+                                     ConstantPoolCacheEntry::flags_offset())));
+      NOT_LP64(__ mov(rbx, rsp));
+      __ shrl(RCX, ConstantPoolCacheEntry::tos_state_shift);
+
+      // Make sure we don't need to mask rcx after the above shift
+      ConstantPoolCacheEntry::verify_tos_state_shift();
+#ifdef _LP64
+      __ movptr(c_rarg1, at_tos_p1());  // initially assume a one word jvalue
+      __ cmpl(c_rarg3, ltos);
+      __ cmovptr(Assembler::equal,
+                 c_rarg1, at_tos_p2()); // ltos (two word jvalue)
+      __ cmpl(c_rarg3, dtos);
+      __ cmovptr(Assembler::equal,
+                 c_rarg1, at_tos_p2()); // dtos (two word jvalue)
+#else
+      __ cmpl(rcx, ltos);
+      __ jccb(Assembler::equal, two_word);
+      __ cmpl(rcx, dtos);
+      __ jccb(Assembler::equal, two_word);
+      __ addptr(rbx, Interpreter::expr_offset_in_bytes(1)); // one word jvalue (not ltos, dtos)
+      __ jmpb(valsize_known);
+
+      __ bind(two_word);
+      __ addptr(rbx, Interpreter::expr_offset_in_bytes(2)); // two words jvalue
+
+      __ bind(valsize_known);
+      // setup object pointer
+      __ movptr(rbx, Address(rbx, 0));
+#endif
+    }
+    // cache entry pointer
+    __ addptr(robj, in_bytes(cp_base_offset));
+    __ shll(RDX, LogBytesPerWord);
+    __ addptr(robj, RDX);
+    // object (tos)
+    __ mov(RCX, rsp);
+    // c_rarg1: object pointer set up above (NULL if static)
+    // c_rarg2: cache entry pointer
+    // c_rarg3: jvalue object on the stack
+    __ call_VM(noreg,
+               CAST_FROM_FN_PTR(address,
+                                InterpreterRuntime::post_field_modification),
+               RBX, robj, RCX);
+    __ get_cache_and_index_at_bcp(cache, index, 1);
+    __ bind(L1);
+  }
+}
+
+void TemplateTable::putfield_or_static(int byte_no, bool is_static) {
+  transition(vtos, vtos);
+
+  const Register cache = rcx;
+  const Register index = rdx;
+  const Register obj   = rcx;
+  const Register off   = rbx;
+  const Register flags = rax;
+  const Register bc    = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+
+  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
+  jvmti_post_field_mod(cache, index, is_static);
+  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
+
+  // [jk] not needed currently
+  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadStore |
+  //                                              Assembler::StoreStore));
+
+  Label notVolatile, Done;
+  __ movl(rdx, flags);
+  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
+  __ andl(rdx, 0x1);
+
+  // field addresses
+  const Address field(obj, off, Address::times_1, 0*wordSize);
+  NOT_LP64( const Address hi(obj, off, Address::times_1, 1*wordSize);)
+
+  Label notByte, notInt, notShort, notChar,
+        notLong, notFloat, notObj, notDouble;
+
+  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
+
+  assert(btos == 0, "change code, btos != 0");
+  __ andl(flags, ConstantPoolCacheEntry::tos_state_mask);
+  __ jcc(Assembler::notZero, notByte);
+
+  // btos
+  {
+    __ pop(btos);
+    if (!is_static) pop_and_check_object(obj);
+    __ movb(field, rax);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_bputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notByte);
+  __ cmpl(flags, atos);
+  __ jcc(Assembler::notEqual, notObj);
+
+  // atos
+  {
+    __ pop(atos);
+    if (!is_static) pop_and_check_object(obj);
+    // Store into the field
+    do_oop_store(_masm, field, rax, _bs->kind(), false);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_aputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notObj);
+  __ cmpl(flags, itos);
+  __ jcc(Assembler::notEqual, notInt);
+
+  // itos
+  {
+    __ pop(itos);
+    if (!is_static) pop_and_check_object(obj);
+    __ movl(field, rax);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_iputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notInt);
+  __ cmpl(flags, ctos);
+  __ jcc(Assembler::notEqual, notChar);
+
+  // ctos
+  {
+    __ pop(ctos);
+    if (!is_static) pop_and_check_object(obj);
+    __ movw(field, rax);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_cputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notChar);
+  __ cmpl(flags, stos);
+  __ jcc(Assembler::notEqual, notShort);
+
+  // stos
+  {
+    __ pop(stos);
+    if (!is_static) pop_and_check_object(obj);
+    __ movw(field, rax);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_sputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notShort);
+  __ cmpl(flags, ltos);
+  __ jcc(Assembler::notEqual, notLong);
+
+  // ltos
+#ifdef _LP64
+  {
+    __ pop(ltos);
+    if (!is_static) pop_and_check_object(obj);
+    __ movq(field, rax);
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_lputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+#else
+  {
+    Label notVolatileLong;
+    __ testl(rdx, rdx);
+    __ jcc(Assembler::zero, notVolatileLong);
+
+    __ pop(ltos);  // overwrites rdx, do this after testing volatile.
+    if (!is_static) pop_and_check_object(obj);
+
+    // Replace with real volatile test
+    __ push(rdx);
+    __ push(rax);                 // Must update atomically with FIST
+    __ fild_d(Address(rsp,0));    // So load into FPU register
+    __ fistp_d(field);            // and put into memory atomically
+    __ addptr(rsp, 2*wordSize);
+    // volatile_barrier();
+    volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
+                                                 Assembler::StoreStore));
+    // Don't rewrite volatile version
+    __ jmp(notVolatile);
+
+    __ bind(notVolatileLong);
+
+    __ pop(ltos);  // overwrites rdx
+    if (!is_static) pop_and_check_object(obj);
+    __ movptr(hi, rdx);
+    __ movptr(field, rax);
+    // Don't rewrite to _fast_lputfield for potential volatile case.
+    __ jmp(notVolatile);
+  }
+#endif // _LP64
+
+  __ bind(notLong);
+  __ cmpl(flags, ftos);
+  __ jcc(Assembler::notEqual, notFloat);
+
+  // ftos
+  {
+    __ pop(ftos);
+    if (!is_static) pop_and_check_object(obj);
+    NOT_LP64( __ fstp_s(field);)
+    LP64_ONLY( __ movflt(field, xmm0);)
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_fputfield, bc, rbx, true, byte_no);
+    }
+    __ jmp(Done);
+  }
+
+  __ bind(notFloat);
+#ifdef ASSERT
+  __ cmpl(flags, dtos);
+  __ jcc(Assembler::notEqual, notDouble);
+#endif
+
+  // dtos
+  {
+    __ pop(dtos);
+    if (!is_static) pop_and_check_object(obj);
+    NOT_LP64( __ fstp_d(field);)
+    LP64_ONLY( __ movdbl(field, xmm0);)
+    if (!is_static) {
+      patch_bytecode(Bytecodes::_fast_dputfield, bc, rbx, true, byte_no);
+    }
+  }
+
+#ifdef ASSERT
+  __ jmp(Done);
+
+  __ bind(notDouble);
+  __ stop("Bad state");
+#endif
+
+  __ bind(Done);
+
+  // Check for volatile store
+  __ testl(rdx, rdx);
+  __ jcc(Assembler::zero, notVolatile);
+  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
+                                               Assembler::StoreStore));
+  __ bind(notVolatile);
+}
+
+void TemplateTable::putfield(int byte_no) {
+  putfield_or_static(byte_no, false);
+}
+
+void TemplateTable::putstatic(int byte_no) {
+  putfield_or_static(byte_no, true);
+}
+
+void TemplateTable::jvmti_post_fast_field_mod() {
+
+  const Register scratch = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+
+  if (JvmtiExport::can_post_field_modification()) {
+    // Check to see if a field modification watch has been set before
+    // we take the time to call into the VM.
+    Label L2;
+    __ mov32(scratch, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
+    __ testl(scratch, scratch);
+    __ jcc(Assembler::zero, L2);
+    __ pop_ptr(rbx);                  // copy the object pointer from tos
+    __ verify_oop(rbx);
+    __ push_ptr(rbx);                 // put the object pointer back on tos
+    // Save tos values before call_VM() clobbers them. Since we have
+    // to do it for every data type, we use the saved values as the
+    // jvalue object.
+    switch (bytecode()) {          // load values into the jvalue object
+    case Bytecodes::_fast_aputfield: __ push_ptr(rax); break;
+    case Bytecodes::_fast_bputfield: // fall through
+    case Bytecodes::_fast_sputfield: // fall through
+    case Bytecodes::_fast_cputfield: // fall through
+    case Bytecodes::_fast_iputfield: __ push_i(rax); break;
+    case Bytecodes::_fast_dputfield: __ push_d(); break;
+    case Bytecodes::_fast_fputfield: __ push_f(); break;
+    case Bytecodes::_fast_lputfield: __ push_l(rax); break;
+
+    default:
+      ShouldNotReachHere();
+    }
+    __ mov(scratch, rsp);             // points to jvalue on the stack
+    // access constant pool cache entry
+    LP64_ONLY(__ get_cache_entry_pointer_at_bcp(c_rarg2, rax, 1));
+    NOT_LP64(__ get_cache_entry_pointer_at_bcp(rax, rdx, 1));
+    __ verify_oop(rbx);
+    // rbx: object pointer copied above
+    // c_rarg2: cache entry pointer
+    // c_rarg3: jvalue object on the stack
+    LP64_ONLY(__ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_modification), rbx, c_rarg2, c_rarg3));
+    NOT_LP64(__ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_modification), rbx, rax, rcx));
+
+    switch (bytecode()) {             // restore tos values
+    case Bytecodes::_fast_aputfield: __ pop_ptr(rax); break;
+    case Bytecodes::_fast_bputfield: // fall through
+    case Bytecodes::_fast_sputfield: // fall through
+    case Bytecodes::_fast_cputfield: // fall through
+    case Bytecodes::_fast_iputfield: __ pop_i(rax); break;
+    case Bytecodes::_fast_dputfield: __ pop_d(); break;
+    case Bytecodes::_fast_fputfield: __ pop_f(); break;
+    case Bytecodes::_fast_lputfield: __ pop_l(rax); break;
+    }
+    __ bind(L2);
+  }
+}
+
+void TemplateTable::fast_storefield(TosState state) {
+  transition(state, vtos);
+
+  ByteSize base = ConstantPoolCache::base_offset();
+
+  jvmti_post_fast_field_mod();
+
+  // access constant pool cache
+  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
+
+  // test for volatile with rdx but rdx is tos register for lputfield.
+  __ movl(rdx, Address(rcx, rbx, Address::times_ptr,
+                       in_bytes(base +
+                                ConstantPoolCacheEntry::flags_offset())));
+
+  // replace index with field offset from cache entry
+  __ movptr(rbx, Address(rcx, rbx, Address::times_ptr,
+                         in_bytes(base + ConstantPoolCacheEntry::f2_offset())));
+
+  // [jk] not needed currently
+  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadStore |
+  //                                              Assembler::StoreStore));
+
+  Label notVolatile;
+  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
+  __ andl(rdx, 0x1);
+
+  // Get object from stack
+  pop_and_check_object(rcx);
+
+  // field address
+  const Address field(rcx, rbx, Address::times_1);
+
+  // access field
+  switch (bytecode()) {
+  case Bytecodes::_fast_aputfield:
+    do_oop_store(_masm, field, rax, _bs->kind(), false);
+    break;
+  case Bytecodes::_fast_lputfield:
+#ifdef _LP64
+  __ movq(field, rax);
+#else
+  __ stop("should not be rewritten");
+#endif
+    break;
+  case Bytecodes::_fast_iputfield:
+    __ movl(field, rax);
+    break;
+  case Bytecodes::_fast_bputfield:
+    __ movb(field, rax);
+    break;
+  case Bytecodes::_fast_sputfield:
+    // fall through
+  case Bytecodes::_fast_cputfield:
+    __ movw(field, rax);
+    break;
+  case Bytecodes::_fast_fputfield:
+    NOT_LP64( __ fstp_s(field); )
+    LP64_ONLY( __ movflt(field, xmm0);)
+    break;
+  case Bytecodes::_fast_dputfield:
+    NOT_LP64( __ fstp_d(field); )
+    LP64_ONLY( __ movdbl(field, xmm0);)
+    break;
+  default:
+    ShouldNotReachHere();
+  }
+
+  // Check for volatile store
+  __ testl(rdx, rdx);
+  __ jcc(Assembler::zero, notVolatile);
+  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
+                                               Assembler::StoreStore));
+  __ bind(notVolatile);
+}
+
+void TemplateTable::fast_accessfield(TosState state) {
+  transition(atos, state);
+
+  // Do the JVMTI work here to avoid disturbing the register state below
+  if (JvmtiExport::can_post_field_access()) {
+    // Check to see if a field access watch has been set before we
+    // take the time to call into the VM.
+    Label L1;
+    __ mov32(rcx, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
+    __ testl(rcx, rcx);
+    __ jcc(Assembler::zero, L1);
+    // access constant pool cache entry
+    LP64_ONLY(__ get_cache_entry_pointer_at_bcp(c_rarg2, rcx, 1));
+    NOT_LP64(__ get_cache_entry_pointer_at_bcp(rcx, rdx, 1));
+    __ verify_oop(rax);
+    __ push_ptr(rax);  // save object pointer before call_VM() clobbers it
+    LP64_ONLY(__ mov(c_rarg1, rax));
+    // c_rarg1: object pointer copied above
+    // c_rarg2: cache entry pointer
+    LP64_ONLY(__ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_access), c_rarg1, c_rarg2));
+    NOT_LP64(__ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_access), rax, rcx));
+    __ pop_ptr(rax); // restore object pointer
+    __ bind(L1);
+  }
+
+  // access constant pool cache
+  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
+  // replace index with field offset from cache entry
+  // [jk] not needed currently
+  // if (os::is_MP()) {
+  //   __ movl(rdx, Address(rcx, rbx, Address::times_8,
+  //                        in_bytes(ConstantPoolCache::base_offset() +
+  //                                 ConstantPoolCacheEntry::flags_offset())));
+  //   __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
+  //   __ andl(rdx, 0x1);
+  // }
+  __ movptr(rbx, Address(rcx, rbx, Address::times_ptr,
+                         in_bytes(ConstantPoolCache::base_offset() +
+                                  ConstantPoolCacheEntry::f2_offset())));
+
+  // rax: object
+  __ verify_oop(rax);
+  __ null_check(rax);
+  Address field(rax, rbx, Address::times_1);
+
+  // access field
+  switch (bytecode()) {
+  case Bytecodes::_fast_agetfield:
+    __ load_heap_oop(rax, field);
+    __ verify_oop(rax);
+    break;
+  case Bytecodes::_fast_lgetfield:
+#ifdef _LP64
+  __ movq(rax, field);
+#else
+  __ stop("should not be rewritten");
+#endif
+    break;
+  case Bytecodes::_fast_igetfield:
+    __ movl(rax, field);
+    break;
+  case Bytecodes::_fast_bgetfield:
+    __ movsbl(rax, field);
+    break;
+  case Bytecodes::_fast_sgetfield:
+    __ load_signed_short(rax, field);
+    break;
+  case Bytecodes::_fast_cgetfield:
+    __ load_unsigned_short(rax, field);
+    break;
+  case Bytecodes::_fast_fgetfield:
+    LP64_ONLY(__ movflt(xmm0, field));
+    NOT_LP64(__ fld_s(field));
+    break;
+  case Bytecodes::_fast_dgetfield:
+    LP64_ONLY(__ movdbl(xmm0, field));
+    NOT_LP64(__ fld_d(field));
+    break;
+  default:
+    ShouldNotReachHere();
+  }
+  // [jk] not needed currently
+  // if (os::is_MP()) {
+  //   Label notVolatile;
+  //   __ testl(rdx, rdx);
+  //   __ jcc(Assembler::zero, notVolatile);
+  //   __ membar(Assembler::LoadLoad);
+  //   __ bind(notVolatile);
+  //};
+}
+
+void TemplateTable::fast_xaccess(TosState state) {
+  transition(vtos, state);
+
+  // get receiver
+  __ movptr(rax, aaddress(0));
+  // access constant pool cache
+  __ get_cache_and_index_at_bcp(rcx, rdx, 2);
+  __ movptr(rbx,
+            Address(rcx, rdx, Address::times_ptr,
+                    in_bytes(ConstantPoolCache::base_offset() +
+                             ConstantPoolCacheEntry::f2_offset())));
+  // make sure exception is reported in correct bcp range (getfield is
+  // next instruction)
+  __ increment(rbcp);
+  __ null_check(rax);
+  const Address field = Address(rax, rbx, Address::times_1, 0*wordSize);
+  switch (state) {
+  case itos:
+    __ movl(rax, field);
+    break;
+  case atos:
+    __ load_heap_oop(rax, field);
+    __ verify_oop(rax);
+    break;
+  case ftos:
+    LP64_ONLY(__ movflt(xmm0, field));
+    NOT_LP64(__ fld_s(field));
+    break;
+  default:
+    ShouldNotReachHere();
+  }
+
+  // [jk] not needed currently
+  // if (os::is_MP()) {
+  //   Label notVolatile;
+  //   __ movl(rdx, Address(rcx, rdx, Address::times_8,
+  //                        in_bytes(ConstantPoolCache::base_offset() +
+  //                                 ConstantPoolCacheEntry::flags_offset())));
+  //   __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
+  //   __ testl(rdx, 0x1);
+  //   __ jcc(Assembler::zero, notVolatile);
+  //   __ membar(Assembler::LoadLoad);
+  //   __ bind(notVolatile);
+  // }
+
+  __ decrement(rbcp);
+}
+
+//-----------------------------------------------------------------------------
+// Calls
+
+void TemplateTable::count_calls(Register method, Register temp) {
+  // implemented elsewhere
+  ShouldNotReachHere();
+}
+
+void TemplateTable::prepare_invoke(int byte_no,
+                                   Register method,  // linked method (or i-klass)
+                                   Register index,   // itable index, MethodType, etc.
+                                   Register recv,    // if caller wants to see it
+                                   Register flags    // if caller wants to test it
+                                   ) {
+  // determine flags
+  const Bytecodes::Code code = bytecode();
+  const bool is_invokeinterface  = code == Bytecodes::_invokeinterface;
+  const bool is_invokedynamic    = code == Bytecodes::_invokedynamic;
+  const bool is_invokehandle     = code == Bytecodes::_invokehandle;
+  const bool is_invokevirtual    = code == Bytecodes::_invokevirtual;
+  const bool is_invokespecial    = code == Bytecodes::_invokespecial;
+  const bool load_receiver       = (recv  != noreg);
+  const bool save_flags          = (flags != noreg);
+  assert(load_receiver == (code != Bytecodes::_invokestatic && code != Bytecodes::_invokedynamic), "");
+  assert(save_flags    == (is_invokeinterface || is_invokevirtual), "need flags for vfinal");
+  assert(flags == noreg || flags == rdx, "");
+  assert(recv  == noreg || recv  == rcx, "");
+
+  // setup registers & access constant pool cache
+  if (recv  == noreg)  recv  = rcx;
+  if (flags == noreg)  flags = rdx;
+  assert_different_registers(method, index, recv, flags);
+
+  // save 'interpreter return address'
+  __ save_bcp();
+
+  load_invoke_cp_cache_entry(byte_no, method, index, flags, is_invokevirtual, false, is_invokedynamic);
+
+  // maybe push appendix to arguments (just before return address)
+  if (is_invokedynamic || is_invokehandle) {
+    Label L_no_push;
+    __ testl(flags, (1 << ConstantPoolCacheEntry::has_appendix_shift));
+    __ jcc(Assembler::zero, L_no_push);
+    // Push the appendix as a trailing parameter.
+    // This must be done before we get the receiver,
+    // since the parameter_size includes it.
+    __ push(rbx);
+    __ mov(rbx, index);
+    assert(ConstantPoolCacheEntry::_indy_resolved_references_appendix_offset == 0, "appendix expected at index+0");
+    __ load_resolved_reference_at_index(index, rbx);
+    __ pop(rbx);
+    __ push(index);  // push appendix (MethodType, CallSite, etc.)
+    __ bind(L_no_push);
+  }
+
+  // load receiver if needed (after appendix is pushed so parameter size is correct)
+  // Note: no return address pushed yet
+  if (load_receiver) {
+    __ movl(recv, flags);
+    __ andl(recv, ConstantPoolCacheEntry::parameter_size_mask);
+    const int no_return_pc_pushed_yet = -1;  // argument slot correction before we push return address
+    const int receiver_is_at_end      = -1;  // back off one slot to get receiver
+    Address recv_addr = __ argument_address(recv, no_return_pc_pushed_yet + receiver_is_at_end);
+    __ movptr(recv, recv_addr);
+    __ verify_oop(recv);
+  }
+
+  if (save_flags) {
+    __ movl(rbcp, flags);
+  }
+
+  // compute return type
+  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
+  // Make sure we don't need to mask flags after the above shift
+  ConstantPoolCacheEntry::verify_tos_state_shift();
+  // load return address
+  {
+    const address table_addr = (address) Interpreter::invoke_return_entry_table_for(code);
+    ExternalAddress table(table_addr);
+    LP64_ONLY(__ lea(rscratch1, table));
+    LP64_ONLY(__ movptr(flags, Address(rscratch1, flags, Address::times_ptr)));
+    NOT_LP64(__ movptr(flags, ArrayAddress(table, Address(noreg, flags, Address::times_ptr))));
+  }
+
+  // push return address
+  __ push(flags);
+
+  // Restore flags value from the constant pool cache, and restore rsi
+  // for later null checks.  r13 is the bytecode pointer
+  if (save_flags) {
+    __ movl(flags, rbcp);
+    __ restore_bcp();
+  }
+}
+
+void TemplateTable::invokevirtual_helper(Register index,
+                                         Register recv,
+                                         Register flags) {
+  // Uses temporary registers rax, rdx
+  assert_different_registers(index, recv, rax, rdx);
+  assert(index == rbx, "");
+  assert(recv  == rcx, "");
+
+  // Test for an invoke of a final method
+  Label notFinal;
+  __ movl(rax, flags);
+  __ andl(rax, (1 << ConstantPoolCacheEntry::is_vfinal_shift));
+  __ jcc(Assembler::zero, notFinal);
+
+  const Register method = index;  // method must be rbx
+  assert(method == rbx,
+         "Method* must be rbx for interpreter calling convention");
+
+  // do the call - the index is actually the method to call
+  // that is, f2 is a vtable index if !is_vfinal, else f2 is a Method*
+
+  // It's final, need a null check here!
+  __ null_check(recv);
+
+  // profile this call
+  __ profile_final_call(rax);
+  __ profile_arguments_type(rax, method, rbcp, true);
+
+  __ jump_from_interpreted(method, rax);
+
+  __ bind(notFinal);
+
+  // get receiver klass
+  __ null_check(recv, oopDesc::klass_offset_in_bytes());
+  __ load_klass(rax, recv);
+
+  // profile this call
+  __ profile_virtual_call(rax, rlocals, rdx);
+  // get target Method* & entry point
+  __ lookup_virtual_method(rax, index, method);
+  __ profile_arguments_type(rdx, method, rbcp, true);
+  __ jump_from_interpreted(method, rdx);
+}
+
+void TemplateTable::invokevirtual(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f2_byte, "use this argument");
+  prepare_invoke(byte_no,
+                 rbx,    // method or vtable index
+                 noreg,  // unused itable index
+                 rcx, rdx); // recv, flags
+
+  // rbx: index
+  // rcx: receiver
+  // rdx: flags
+
+  invokevirtual_helper(rbx, rcx, rdx);
+}
+
+void TemplateTable::invokespecial(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f1_byte, "use this argument");
+  prepare_invoke(byte_no, rbx, noreg,  // get f1 Method*
+                 rcx);  // get receiver also for null check
+  __ verify_oop(rcx);
+  __ null_check(rcx);
+  // do the call
+  __ profile_call(rax);
+  __ profile_arguments_type(rax, rbx, rbcp, false);
+  __ jump_from_interpreted(rbx, rax);
+}
+
+void TemplateTable::invokestatic(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f1_byte, "use this argument");
+  prepare_invoke(byte_no, rbx);  // get f1 Method*
+  // do the call
+  __ profile_call(rax);
+  __ profile_arguments_type(rax, rbx, rbcp, false);
+  __ jump_from_interpreted(rbx, rax);
+}
+
+
+void TemplateTable::fast_invokevfinal(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f2_byte, "use this argument");
+  __ stop("fast_invokevfinal not used on x86");
+}
+
+
+void TemplateTable::invokeinterface(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f1_byte, "use this argument");
+  prepare_invoke(byte_no, rax, rbx,  // get f1 Klass*, f2 itable index
+                 rcx, rdx); // recv, flags
+
+  // rax: interface klass (from f1)
+  // rbx: itable index (from f2)
+  // rcx: receiver
+  // rdx: flags
+
+  // Special case of invokeinterface called for virtual method of
+  // java.lang.Object.  See cpCacheOop.cpp for details.
+  // This code isn't produced by javac, but could be produced by
+  // another compliant java compiler.
+  Label notMethod;
+  __ movl(rlocals, rdx);
+  __ andl(rlocals, (1 << ConstantPoolCacheEntry::is_forced_virtual_shift));
+
+  __ jcc(Assembler::zero, notMethod);
+
+  invokevirtual_helper(rbx, rcx, rdx);
+  __ bind(notMethod);
+
+  // Get receiver klass into rdx - also a null check
+  __ restore_locals();  // restore r14
+  __ null_check(rcx, oopDesc::klass_offset_in_bytes());
+  __ load_klass(rdx, rcx);
+
+  // profile this call
+  __ profile_virtual_call(rdx, rbcp, rlocals);
+
+  Label no_such_interface, no_such_method;
+
+  __ lookup_interface_method(// inputs: rec. class, interface, itable index
+                             rdx, rax, rbx,
+                             // outputs: method, scan temp. reg
+                             rbx, rbcp,
+                             no_such_interface);
+
+  // rbx: Method* to call
+  // rcx: receiver
+  // Check for abstract method error
+  // Note: This should be done more efficiently via a throw_abstract_method_error
+  //       interpreter entry point and a conditional jump to it in case of a null
+  //       method.
+  __ testptr(rbx, rbx);
+  __ jcc(Assembler::zero, no_such_method);
+
+  __ profile_arguments_type(rdx, rbx, rbcp, true);
+
+  // do the call
+  // rcx: receiver
+  // rbx,: Method*
+  __ jump_from_interpreted(rbx, rdx);
+  __ should_not_reach_here();
+
+  // exception handling code follows...
+  // note: must restore interpreter registers to canonical
+  //       state for exception handling to work correctly!
+
+  __ bind(no_such_method);
+  // throw exception
+  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
+  __ restore_bcp();      // rbcp must be correct for exception handler   (was destroyed)
+  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
+  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::throw_AbstractMethodError));
+  // the call_VM checks for exception, so we should never return here.
+  __ should_not_reach_here();
+
+  __ bind(no_such_interface);
+  // throw exception
+  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
+  __ restore_bcp();      // rbcp must be correct for exception handler   (was destroyed)
+  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
+  __ call_VM(noreg, CAST_FROM_FN_PTR(address,
+                   InterpreterRuntime::throw_IncompatibleClassChangeError));
+  // the call_VM checks for exception, so we should never return here.
+  __ should_not_reach_here();
+}
+
+void TemplateTable::invokehandle(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f1_byte, "use this argument");
+  const Register rbx_method = rbx;
+  const Register rax_mtype  = rax;
+  const Register rcx_recv   = rcx;
+  const Register rdx_flags  = rdx;
+
+  prepare_invoke(byte_no, rbx_method, rax_mtype, rcx_recv);
+  __ verify_method_ptr(rbx_method);
+  __ verify_oop(rcx_recv);
+  __ null_check(rcx_recv);
+
+  // rax: MethodType object (from cpool->resolved_references[f1], if necessary)
+  // rbx: MH.invokeExact_MT method (from f2)
+
+  // Note:  rax_mtype is already pushed (if necessary) by prepare_invoke
+
+  // FIXME: profile the LambdaForm also
+  __ profile_final_call(rax);
+  __ profile_arguments_type(rdx, rbx_method, rbcp, true);
+
+  __ jump_from_interpreted(rbx_method, rdx);
+}
+
+void TemplateTable::invokedynamic(int byte_no) {
+  transition(vtos, vtos);
+  assert(byte_no == f1_byte, "use this argument");
+
+  const Register rbx_method   = rbx;
+  const Register rax_callsite = rax;
+
+  prepare_invoke(byte_no, rbx_method, rax_callsite);
+
+  // rax: CallSite object (from cpool->resolved_references[f1])
+  // rbx: MH.linkToCallSite method (from f2)
+
+  // Note:  rax_callsite is already pushed by prepare_invoke
+
+  // %%% should make a type profile for any invokedynamic that takes a ref argument
+  // profile this call
+  __ profile_call(rbcp);
+  __ profile_arguments_type(rdx, rbx_method, rbcp, false);
+
+  __ verify_oop(rax_callsite);
+
+  __ jump_from_interpreted(rbx_method, rdx);
+}
+
+//-----------------------------------------------------------------------------
+// Allocation
+
+void TemplateTable::_new() {
+  transition(vtos, atos);
+  __ get_unsigned_2_byte_index_at_bcp(rdx, 1);
+  Label slow_case;
+  Label slow_case_no_pop;
+  Label done;
+  Label initialize_header;
+  Label initialize_object;  // including clearing the fields
+  Label allocate_shared;
+
+  __ get_cpool_and_tags(rcx, rax);
+
+  // Make sure the class we're about to instantiate has been resolved.
+  // This is done before loading InstanceKlass to be consistent with the order
+  // how Constant Pool is updated (see ConstantPool::klass_at_put)
+  const int tags_offset = Array<u1>::base_offset_in_bytes();
+  __ cmpb(Address(rax, rdx, Address::times_1, tags_offset), JVM_CONSTANT_Class);
+  __ jcc(Assembler::notEqual, slow_case_no_pop);
+
+  // get InstanceKlass
+  __ movptr(rcx, Address(rcx, rdx, Address::times_ptr, sizeof(ConstantPool)));
+  __ push(rcx);  // save the contexts of klass for initializing the header
+
+  // make sure klass is initialized & doesn't have finalizer
+  // make sure klass is fully initialized
+  __ cmpb(Address(rcx, InstanceKlass::init_state_offset()), InstanceKlass::fully_initialized);
+  __ jcc(Assembler::notEqual, slow_case);
+
+  // get instance_size in InstanceKlass (scaled to a count of bytes)
+  __ movl(rdx, Address(rcx, Klass::layout_helper_offset()));
+  // test to see if it has a finalizer or is malformed in some way
+  __ testl(rdx, Klass::_lh_instance_slow_path_bit);
+  __ jcc(Assembler::notZero, slow_case);
+
+  //
+  // Allocate the instance
+  // 1) Try to allocate in the TLAB
+  // 2) if fail and the object is large allocate in the shared Eden
+  // 3) if the above fails (or is not applicable), go to a slow case
+  // (creates a new TLAB, etc.)
+
+  const bool allow_shared_alloc =
+    Universe::heap()->supports_inline_contig_alloc();
+
+  const Register thread = LP64_ONLY(r15_thread) NOT_LP64(rcx);
+#ifndef _LP64
+  if (UseTLAB || allow_shared_alloc) {
+    __ get_thread(thread);
+  }
+#endif // _LP64
+
+  if (UseTLAB) {
+    __ movptr(rax, Address(thread, in_bytes(JavaThread::tlab_top_offset())));
+    __ lea(rbx, Address(rax, rdx, Address::times_1));
+    __ cmpptr(rbx, Address(thread, in_bytes(JavaThread::tlab_end_offset())));
+    __ jcc(Assembler::above, allow_shared_alloc ? allocate_shared : slow_case);
+    __ movptr(Address(thread, in_bytes(JavaThread::tlab_top_offset())), rbx);
+    if (ZeroTLAB) {
+      // the fields have been already cleared
+      __ jmp(initialize_header);
+    } else {
+      // initialize both the header and fields
+      __ jmp(initialize_object);
+    }
+  }
+
+  // Allocation in the shared Eden, if allowed.
+  //
+  // rdx: instance size in bytes
+  if (allow_shared_alloc) {
+    __ bind(allocate_shared);
+
+    ExternalAddress heap_top((address)Universe::heap()->top_addr());
+    ExternalAddress heap_end((address)Universe::heap()->end_addr());
+
+    Label retry;
+    __ bind(retry);
+    __ movptr(rax, heap_top);
+    __ lea(rbx, Address(rax, rdx, Address::times_1));
+    __ cmpptr(rbx, heap_end);
+    __ jcc(Assembler::above, slow_case);
+
+    // Compare rax, with the top addr, and if still equal, store the new
+    // top addr in rbx, at the address of the top addr pointer. Sets ZF if was
+    // equal, and clears it otherwise. Use lock prefix for atomicity on MPs.
+    //
+    // rax,: object begin
+    // rbx,: object end
+    // rdx: instance size in bytes
+    __ locked_cmpxchgptr(rbx, heap_top);
+
+    // if someone beat us on the allocation, try again, otherwise continue
+    __ jcc(Assembler::notEqual, retry);
+
+    __ incr_allocated_bytes(thread, rdx, 0);
+  }
+
+  if (UseTLAB || Universe::heap()->supports_inline_contig_alloc()) {
+    // The object is initialized before the header.  If the object size is
+    // zero, go directly to the header initialization.
+    __ bind(initialize_object);
+    __ decrement(rdx, sizeof(oopDesc));
+    __ jcc(Assembler::zero, initialize_header);
+
+    // Initialize topmost object field, divide rdx by 8, check if odd and
+    // test if zero.
+    __ xorl(rcx, rcx);    // use zero reg to clear memory (shorter code)
+    __ shrl(rdx, LogBytesPerLong); // divide by 2*oopSize and set carry flag if odd
+
+    // rdx must have been multiple of 8
+#ifdef ASSERT
+    // make sure rdx was multiple of 8
+    Label L;
+    // Ignore partial flag stall after shrl() since it is debug VM
+    __ jccb(Assembler::carryClear, L);
+    __ stop("object size is not multiple of 2 - adjust this code");
+    __ bind(L);
+    // rdx must be > 0, no extra check needed here
+#endif
+
+    // initialize remaining object fields: rdx was a multiple of 8
+    { Label loop;
+    __ bind(loop);
+    __ movptr(Address(rax, rdx, Address::times_8, sizeof(oopDesc) - 1*oopSize), rcx);
+    NOT_LP64(__ movptr(Address(rax, rdx, Address::times_8, sizeof(oopDesc) - 2*oopSize), rcx));
+    __ decrement(rdx);
+    __ jcc(Assembler::notZero, loop);
+    }
+
+    // initialize object header only.
+    __ bind(initialize_header);
+    if (UseBiasedLocking) {
+      __ pop(rcx);   // get saved klass back in the register.
+      __ movptr(rbx, Address(rcx, Klass::prototype_header_offset()));
+      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes ()), rbx);
+    } else {
+      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes ()),
+                (intptr_t)markOopDesc::prototype()); // header
+      __ pop(rcx);   // get saved klass back in the register.
+    }
+#ifdef _LP64
+    __ xorl(rsi, rsi); // use zero reg to clear memory (shorter code)
+    __ store_klass_gap(rax, rsi);  // zero klass gap for compressed oops
+#endif
+    __ store_klass(rax, rcx);  // klass
+
+    {
+      SkipIfEqual skip_if(_masm, &DTraceAllocProbes, 0);
+      // Trigger dtrace event for fastpath
+      __ push(atos);
+      __ call_VM_leaf(
+           CAST_FROM_FN_PTR(address, SharedRuntime::dtrace_object_alloc), rax);
+      __ pop(atos);
+    }
+
+    __ jmp(done);
+  }
+
+  // slow case
+  __ bind(slow_case);
+  __ pop(rcx);   // restore stack pointer to what it was when we came in.
+  __ bind(slow_case_no_pop);
+
+  Register rarg1 = LP64_ONLY(c_rarg1) NOT_LP64(rax);
+  Register rarg2 = LP64_ONLY(c_rarg2) NOT_LP64(rdx);
+
+  __ get_constant_pool(rarg1);
+  __ get_unsigned_2_byte_index_at_bcp(rarg2, 1);
+  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::_new), rarg1, rarg2);
+   __ verify_oop(rax);
+
+  // continue
+  __ bind(done);
+}
+
+void TemplateTable::newarray() {
+  transition(itos, atos);
+  Register rarg1 = LP64_ONLY(c_rarg1) NOT_LP64(rdx);
+  __ load_unsigned_byte(rarg1, at_bcp(1));
+  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::newarray),
+          rarg1, rax);
+}
+
+void TemplateTable::anewarray() {
+  transition(itos, atos);
+
+  Register rarg1 = LP64_ONLY(c_rarg1) NOT_LP64(rcx);
+  Register rarg2 = LP64_ONLY(c_rarg2) NOT_LP64(rdx);
+
+  __ get_unsigned_2_byte_index_at_bcp(rarg2, 1);
+  __ get_constant_pool(rarg1);
+  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::anewarray),
+          rarg1, rarg2, rax);
+}
+
+void TemplateTable::arraylength() {
+  transition(atos, itos);
+  __ null_check(rax, arrayOopDesc::length_offset_in_bytes());
+  __ movl(rax, Address(rax, arrayOopDesc::length_offset_in_bytes()));
+}
+
+void TemplateTable::checkcast() {
+  transition(atos, atos);
+  Label done, is_null, ok_is_subtype, quicked, resolved;
+  __ testptr(rax, rax); // object is in rax
+  __ jcc(Assembler::zero, is_null);
+
+  // Get cpool & tags index
+  __ get_cpool_and_tags(rcx, rdx); // rcx=cpool, rdx=tags array
+  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // rbx=index
+  // See if bytecode has already been quicked
+  __ cmpb(Address(rdx, rbx,
+                  Address::times_1,
+                  Array<u1>::base_offset_in_bytes()),
+          JVM_CONSTANT_Class);
+  __ jcc(Assembler::equal, quicked);
+  __ push(atos); // save receiver for result, and for GC
+  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc));
+
+  // vm_result_2 has metadata result
+#ifndef _LP64
+  // borrow rdi from locals
+  __ get_thread(rdi);
+  __ get_vm_result_2(rax, rdi);
+  __ restore_locals();
+#else
+  __ get_vm_result_2(rax, r15_thread);
+#endif
+
+  __ pop_ptr(rdx); // restore receiver
+  __ jmpb(resolved);
+
+  // Get superklass in rax and subklass in rbx
+  __ bind(quicked);
+  __ mov(rdx, rax); // Save object in rdx; rax needed for subtype check
+  __ movptr(rax, Address(rcx, rbx,
+                       Address::times_ptr, sizeof(ConstantPool)));
+
+  __ bind(resolved);
+  __ load_klass(rbx, rdx);
+
+  // Generate subtype check.  Blows rcx, rdi.  Object in rdx.
+  // Superklass in rax.  Subklass in rbx.
+  __ gen_subtype_check(rbx, ok_is_subtype);
+
+  // Come here on failure
+  __ push_ptr(rdx);
+  // object is at TOS
+  __ jump(ExternalAddress(Interpreter::_throw_ClassCastException_entry));
+
+  // Come here on success
+  __ bind(ok_is_subtype);
+  __ mov(rax, rdx); // Restore object in rdx
+
+  // Collect counts on whether this check-cast sees NULLs a lot or not.
+  if (ProfileInterpreter) {
+    __ jmp(done);
+    __ bind(is_null);
+    __ profile_null_seen(rcx);
+  } else {
+    __ bind(is_null);   // same as 'done'
+  }
+  __ bind(done);
+}
+
+void TemplateTable::instanceof() {
+  transition(atos, itos);
+  Label done, is_null, ok_is_subtype, quicked, resolved;
+  __ testptr(rax, rax);
+  __ jcc(Assembler::zero, is_null);
+
+  // Get cpool & tags index
+  __ get_cpool_and_tags(rcx, rdx); // rcx=cpool, rdx=tags array
+  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // rbx=index
+  // See if bytecode has already been quicked
+  __ cmpb(Address(rdx, rbx,
+                  Address::times_1,
+                  Array<u1>::base_offset_in_bytes()),
+          JVM_CONSTANT_Class);
+  __ jcc(Assembler::equal, quicked);
+
+  __ push(atos); // save receiver for result, and for GC
+  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc));
+  // vm_result_2 has metadata result
+
+#ifndef _LP64
+  // borrow rdi from locals
+  __ get_thread(rdi);
+  __ get_vm_result_2(rax, rdi);
+  __ restore_locals();
+#else
+  __ get_vm_result_2(rax, r15_thread);
+#endif
+
+  __ pop_ptr(rdx); // restore receiver
+  __ verify_oop(rdx);
+  __ load_klass(rdx, rdx);
+  __ jmpb(resolved);
+
+  // Get superklass in rax and subklass in rdx
+  __ bind(quicked);
+  __ load_klass(rdx, rax);
+  __ movptr(rax, Address(rcx, rbx,
+                         Address::times_ptr, sizeof(ConstantPool)));
+
+  __ bind(resolved);
+
+  // Generate subtype check.  Blows rcx, rdi
+  // Superklass in rax.  Subklass in rdx.
+  __ gen_subtype_check(rdx, ok_is_subtype);
+
+  // Come here on failure
+  __ xorl(rax, rax);
+  __ jmpb(done);
+  // Come here on success
+  __ bind(ok_is_subtype);
+  __ movl(rax, 1);
+
+  // Collect counts on whether this test sees NULLs a lot or not.
+  if (ProfileInterpreter) {
+    __ jmp(done);
+    __ bind(is_null);
+    __ profile_null_seen(rcx);
+  } else {
+    __ bind(is_null);   // same as 'done'
+  }
+  __ bind(done);
+  // rax = 0: obj == NULL or  obj is not an instanceof the specified klass
+  // rax = 1: obj != NULL and obj is     an instanceof the specified klass
+}
+
+
+//----------------------------------------------------------------------------------------------------
+// Breakpoints
+void TemplateTable::_breakpoint() {
+  // Note: We get here even if we are single stepping..
+  // jbug insists on setting breakpoints at every bytecode
+  // even if we are in single step mode.
+
+  transition(vtos, vtos);
+
+  Register rarg = LP64_ONLY(c_rarg1) NOT_LP64(rcx);
+
+  // get the unpatched byte code
+  __ get_method(rarg);
+  __ call_VM(noreg,
+             CAST_FROM_FN_PTR(address,
+                              InterpreterRuntime::get_original_bytecode_at),
+             rarg, rbcp);
+  __ mov(rbx, rax);  // why?
+
+  // post the breakpoint event
+  __ get_method(rarg);
+  __ call_VM(noreg,
+             CAST_FROM_FN_PTR(address, InterpreterRuntime::_breakpoint),
+             rarg, rbcp);
+
+  // complete the execution of original bytecode
+  __ dispatch_only_normal(vtos);
+}
+
+//-----------------------------------------------------------------------------
+// Exceptions
+
+void TemplateTable::athrow() {
+  transition(atos, vtos);
+  __ null_check(rax);
+  __ jump(ExternalAddress(Interpreter::throw_exception_entry()));
+}
+
+//-----------------------------------------------------------------------------
+// Synchronization
+//
+// Note: monitorenter & exit are symmetric routines; which is reflected
+//       in the assembly code structure as well
+//
+// Stack layout:
+//
+// [expressions  ] <--- rsp               = expression stack top
+// ..
+// [expressions  ]
+// [monitor entry] <--- monitor block top = expression stack bot
+// ..
+// [monitor entry]
+// [frame data   ] <--- monitor block bot
+// ...
+// [saved rbp    ] <--- rbp
+void TemplateTable::monitorenter() {
+  transition(atos, vtos);
+
+  // check for NULL object
+  __ null_check(rax);
+
+  const Address monitor_block_top(
+        rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
+  const Address monitor_block_bot(
+        rbp, frame::interpreter_frame_initial_sp_offset * wordSize);
+  const int entry_size = frame::interpreter_frame_monitor_size() * wordSize;
+
+  Label allocated;
+
+  Register rtop = LP64_ONLY(c_rarg3) NOT_LP64(rcx);
+  Register rbot = LP64_ONLY(c_rarg2) NOT_LP64(rbx);
+  Register rmon = LP64_ONLY(c_rarg1) NOT_LP64(rdx);
+
+  // initialize entry pointer
+  __ xorl(rmon, rmon); // points to free slot or NULL
+
+  // find a free slot in the monitor block (result in rmon)
+  {
+    Label entry, loop, exit;
+    __ movptr(rtop, monitor_block_top); // points to current entry,
+                                        // starting with top-most entry
+    __ lea(rbot, monitor_block_bot);    // points to word before bottom
+                                        // of monitor block
+    __ jmpb(entry);
+
+    __ bind(loop);
+    // check if current entry is used
+    __ cmpptr(Address(rtop, BasicObjectLock::obj_offset_in_bytes()), (int32_t) NULL_WORD);
+    // if not used then remember entry in rmon
+    __ cmovptr(Assembler::equal, rmon, rtop);   // cmov => cmovptr
+    // check if current entry is for same object
+    __ cmpptr(rax, Address(rtop, BasicObjectLock::obj_offset_in_bytes()));
+    // if same object then stop searching
+    __ jccb(Assembler::equal, exit);
+    // otherwise advance to next entry
+    __ addptr(rtop, entry_size);
+    __ bind(entry);
+    // check if bottom reached
+    __ cmpptr(rtop, rbot);
+    // if not at bottom then check this entry
+    __ jcc(Assembler::notEqual, loop);
+    __ bind(exit);
+  }
+
+  __ testptr(rmon, rmon); // check if a slot has been found
+  __ jcc(Assembler::notZero, allocated); // if found, continue with that one
+
+  // allocate one if there's no free slot
+  {
+    Label entry, loop;
+    // 1. compute new pointers          // rsp: old expression stack top
+    __ movptr(rmon, monitor_block_bot); // rmon: old expression stack bottom
+    __ subptr(rsp, entry_size);         // move expression stack top
+    __ subptr(rmon, entry_size);        // move expression stack bottom
+    __ mov(rtop, rsp);                  // set start value for copy loop
+    __ movptr(monitor_block_bot, rmon); // set new monitor block bottom
+    __ jmp(entry);
+    // 2. move expression stack contents
+    __ bind(loop);
+    __ movptr(rbot, Address(rtop, entry_size)); // load expression stack
+                                                // word from old location
+    __ movptr(Address(rtop, 0), rbot);          // and store it at new location
+    __ addptr(rtop, wordSize);                  // advance to next word
+    __ bind(entry);
+    __ cmpptr(rtop, rmon);                      // check if bottom reached
+    __ jcc(Assembler::notEqual, loop);          // if not at bottom then
+                                                // copy next word
+  }
+
+  // call run-time routine
+  // rmon: points to monitor entry
+  __ bind(allocated);
+
+  // Increment bcp to point to the next bytecode, so exception
+  // handling for async. exceptions work correctly.
+  // The object has already been poped from the stack, so the
+  // expression stack looks correct.
+  __ increment(rbcp);
+
+  // store object
+  __ movptr(Address(rmon, BasicObjectLock::obj_offset_in_bytes()), rax);
+  __ lock_object(rmon);
+
+  // check to make sure this monitor doesn't cause stack overflow after locking
+  __ save_bcp();  // in case of exception
+  __ generate_stack_overflow_check(0);
+
+  // The bcp has already been incremented. Just need to dispatch to
+  // next instruction.
+  __ dispatch_next(vtos);
+}
+
+void TemplateTable::monitorexit() {
+  transition(atos, vtos);
+
+  // check for NULL object
+  __ null_check(rax);
+
+  const Address monitor_block_top(
+        rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
+  const Address monitor_block_bot(
+        rbp, frame::interpreter_frame_initial_sp_offset * wordSize);
+  const int entry_size = frame::interpreter_frame_monitor_size() * wordSize;
+
+  Register rtop = LP64_ONLY(c_rarg1) NOT_LP64(rdx);
+  Register rbot = LP64_ONLY(c_rarg2) NOT_LP64(rbx);
+
+  Label found;
+
+  // find matching slot
+  {
+    Label entry, loop;
+    __ movptr(rtop, monitor_block_top); // points to current entry,
+                                        // starting with top-most entry
+    __ lea(rbot, monitor_block_bot);    // points to word before bottom
+                                        // of monitor block
+    __ jmpb(entry);
+
+    __ bind(loop);
+    // check if current entry is for same object
+    __ cmpptr(rax, Address(rtop, BasicObjectLock::obj_offset_in_bytes()));
+    // if same object then stop searching
+    __ jcc(Assembler::equal, found);
+    // otherwise advance to next entry
+    __ addptr(rtop, entry_size);
+    __ bind(entry);
+    // check if bottom reached
+    __ cmpptr(rtop, rbot);
+    // if not at bottom then check this entry
+    __ jcc(Assembler::notEqual, loop);
+  }
+
+  // error handling. Unlocking was not block-structured
+  __ call_VM(noreg, CAST_FROM_FN_PTR(address,
+                   InterpreterRuntime::throw_illegal_monitor_state_exception));
+  __ should_not_reach_here();
+
+  // call run-time routine
+  __ bind(found);
+  __ push_ptr(rax); // make sure object is on stack (contract with oopMaps)
+  __ unlock_object(rtop);
+  __ pop_ptr(rax); // discard object
+}
+
+// Wide instructions
+void TemplateTable::wide() {
+  transition(vtos, vtos);
+  __ load_unsigned_byte(rbx, at_bcp(1));
+  ExternalAddress wtable((address)Interpreter::_wentry_point);
+  __ jump(ArrayAddress(wtable, Address(noreg, rbx, Address::times_ptr)));
+  // Note: the rbcp increment step is part of the individual wide bytecode implementations
+}
+
+// Multi arrays
+void TemplateTable::multianewarray() {
+  transition(vtos, atos);
+
+  Register rarg = LP64_ONLY(c_rarg1) NOT_LP64(rax);
+  __ load_unsigned_byte(rax, at_bcp(3)); // get number of dimensions
+  // last dim is on top of stack; we want address of first one:
+  // first_addr = last_addr + (ndims - 1) * stackElementSize - 1*wordsize
+  // the latter wordSize to point to the beginning of the array.
+  __ lea(rarg, Address(rsp, rax, Interpreter::stackElementScale(), -wordSize));
+  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::multianewarray), rarg);
+  __ load_unsigned_byte(rbx, at_bcp(3));
+  __ lea(rsp, Address(rsp, rbx, Interpreter::stackElementScale()));  // get rid of counts
+}
+#endif /* !CC_INTERP */
+
--- a/hotspot/src/cpu/x86/vm/templateTable_x86_32.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,3668 +0,0 @@
-/*
- * Copyright (c) 1997, 2014, 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 "asm/macroAssembler.hpp"
-#include "interpreter/interpreter.hpp"
-#include "interpreter/interpreterRuntime.hpp"
-#include "interpreter/interp_masm.hpp"
-#include "interpreter/templateTable.hpp"
-#include "memory/universe.inline.hpp"
-#include "oops/methodData.hpp"
-#include "oops/objArrayKlass.hpp"
-#include "oops/oop.inline.hpp"
-#include "prims/methodHandles.hpp"
-#include "runtime/sharedRuntime.hpp"
-#include "runtime/stubRoutines.hpp"
-#include "runtime/synchronizer.hpp"
-#include "utilities/macros.hpp"
-
-#ifndef CC_INTERP
-#define __ _masm->
-
-//----------------------------------------------------------------------------------------------------
-// Platform-dependent initialization
-
-void TemplateTable::pd_initialize() {
-  // No i486 specific initialization
-}
-
-//----------------------------------------------------------------------------------------------------
-// Address computation
-
-// local variables
-static inline Address iaddress(int n)            {
-  return Address(rdi, Interpreter::local_offset_in_bytes(n));
-}
-
-static inline Address laddress(int n)            { return iaddress(n + 1); }
-static inline Address haddress(int n)            { return iaddress(n + 0); }
-static inline Address faddress(int n)            { return iaddress(n); }
-static inline Address daddress(int n)            { return laddress(n); }
-static inline Address aaddress(int n)            { return iaddress(n); }
-
-static inline Address iaddress(Register r)       {
-  return Address(rdi, r, Interpreter::stackElementScale());
-}
-static inline Address laddress(Register r)       {
-  return Address(rdi, r, Interpreter::stackElementScale(), Interpreter::local_offset_in_bytes(1));
-}
-static inline Address haddress(Register r)       {
-  return Address(rdi, r, Interpreter::stackElementScale(), Interpreter::local_offset_in_bytes(0));
-}
-
-static inline Address faddress(Register r)       { return iaddress(r); }
-static inline Address daddress(Register r)       { return laddress(r); }
-static inline Address aaddress(Register r)       { return iaddress(r); }
-
-// expression stack
-// (Note: Must not use symmetric equivalents at_rsp_m1/2 since they store
-// data beyond the rsp which is potentially unsafe in an MT environment;
-// an interrupt may overwrite that data.)
-static inline Address at_rsp   () {
-  return Address(rsp, 0);
-}
-
-// At top of Java expression stack which may be different than rsp().  It
-// isn't for category 1 objects.
-static inline Address at_tos   () {
-  Address tos = Address(rsp,  Interpreter::expr_offset_in_bytes(0));
-  return tos;
-}
-
-static inline Address at_tos_p1() {
-  return Address(rsp,  Interpreter::expr_offset_in_bytes(1));
-}
-
-static inline Address at_tos_p2() {
-  return Address(rsp,  Interpreter::expr_offset_in_bytes(2));
-}
-
-// Condition conversion
-static Assembler::Condition j_not(TemplateTable::Condition cc) {
-  switch (cc) {
-    case TemplateTable::equal        : return Assembler::notEqual;
-    case TemplateTable::not_equal    : return Assembler::equal;
-    case TemplateTable::less         : return Assembler::greaterEqual;
-    case TemplateTable::less_equal   : return Assembler::greater;
-    case TemplateTable::greater      : return Assembler::lessEqual;
-    case TemplateTable::greater_equal: return Assembler::less;
-  }
-  ShouldNotReachHere();
-  return Assembler::zero;
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Miscelaneous helper routines
-
-// Store an oop (or NULL) at the address described by obj.
-// If val == noreg this means store a NULL
-
-static void do_oop_store(InterpreterMacroAssembler* _masm,
-                         Address obj,
-                         Register val,
-                         BarrierSet::Name barrier,
-                         bool precise) {
-  assert(val == noreg || val == rax, "parameter is just for looks");
-  switch (barrier) {
-#if INCLUDE_ALL_GCS
-    case BarrierSet::G1SATBCT:
-    case BarrierSet::G1SATBCTLogging:
-      {
-        // flatten object address if needed
-        // We do it regardless of precise because we need the registers
-        if (obj.index() == noreg && obj.disp() == 0) {
-          if (obj.base() != rdx) {
-            __ movl(rdx, obj.base());
-          }
-        } else {
-          __ leal(rdx, obj);
-        }
-        __ get_thread(rcx);
-        __ save_bcp();
-        __ g1_write_barrier_pre(rdx /* obj */,
-                                rbx /* pre_val */,
-                                rcx /* thread */,
-                                rsi /* tmp */,
-                                val != noreg /* tosca_live */,
-                                false /* expand_call */);
-
-        // Do the actual store
-        // noreg means NULL
-        if (val == noreg) {
-          __ movptr(Address(rdx, 0), NULL_WORD);
-          // No post barrier for NULL
-        } else {
-          __ movl(Address(rdx, 0), val);
-          __ g1_write_barrier_post(rdx /* store_adr */,
-                                   val /* new_val */,
-                                   rcx /* thread */,
-                                   rbx /* tmp */,
-                                   rsi /* tmp2 */);
-        }
-        __ restore_bcp();
-
-      }
-      break;
-#endif // INCLUDE_ALL_GCS
-    case BarrierSet::CardTableModRef:
-    case BarrierSet::CardTableExtension:
-      {
-        if (val == noreg) {
-          __ movptr(obj, NULL_WORD);
-        } else {
-          __ movl(obj, val);
-          // flatten object address if needed
-          if (!precise || (obj.index() == noreg && obj.disp() == 0)) {
-            __ store_check(obj.base());
-          } else {
-            __ leal(rdx, obj);
-            __ store_check(rdx);
-          }
-        }
-      }
-      break;
-    case BarrierSet::ModRef:
-      if (val == noreg) {
-        __ movptr(obj, NULL_WORD);
-      } else {
-        __ movl(obj, val);
-      }
-      break;
-    default      :
-      ShouldNotReachHere();
-
-  }
-}
-
-Address TemplateTable::at_bcp(int offset) {
-  assert(_desc->uses_bcp(), "inconsistent uses_bcp information");
-  return Address(rsi, offset);
-}
-
-
-void TemplateTable::patch_bytecode(Bytecodes::Code bc, Register bc_reg,
-                                   Register temp_reg, bool load_bc_into_bc_reg/*=true*/,
-                                   int byte_no) {
-  if (!RewriteBytecodes)  return;
-  Label L_patch_done;
-
-  switch (bc) {
-  case Bytecodes::_fast_aputfield:
-  case Bytecodes::_fast_bputfield:
-  case Bytecodes::_fast_cputfield:
-  case Bytecodes::_fast_dputfield:
-  case Bytecodes::_fast_fputfield:
-  case Bytecodes::_fast_iputfield:
-  case Bytecodes::_fast_lputfield:
-  case Bytecodes::_fast_sputfield:
-    {
-      // We skip bytecode quickening for putfield instructions when
-      // the put_code written to the constant pool cache is zero.
-      // This is required so that every execution of this instruction
-      // calls out to InterpreterRuntime::resolve_get_put to do
-      // additional, required work.
-      assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
-      assert(load_bc_into_bc_reg, "we use bc_reg as temp");
-      __ get_cache_and_index_and_bytecode_at_bcp(bc_reg, temp_reg, temp_reg, byte_no, 1);
-      __ movl(bc_reg, bc);
-      __ cmpl(temp_reg, (int) 0);
-      __ jcc(Assembler::zero, L_patch_done);  // don't patch
-    }
-    break;
-  default:
-    assert(byte_no == -1, "sanity");
-    // the pair bytecodes have already done the load.
-    if (load_bc_into_bc_reg) {
-      __ movl(bc_reg, bc);
-    }
-  }
-
-  if (JvmtiExport::can_post_breakpoint()) {
-    Label L_fast_patch;
-    // if a breakpoint is present we can't rewrite the stream directly
-    __ movzbl(temp_reg, at_bcp(0));
-    __ cmpl(temp_reg, Bytecodes::_breakpoint);
-    __ jcc(Assembler::notEqual, L_fast_patch);
-    __ get_method(temp_reg);
-    // Let breakpoint table handling rewrite to quicker bytecode
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::set_original_bytecode_at), temp_reg, rsi, bc_reg);
-#ifndef ASSERT
-    __ jmpb(L_patch_done);
-#else
-    __ jmp(L_patch_done);
-#endif
-    __ bind(L_fast_patch);
-  }
-
-#ifdef ASSERT
-  Label L_okay;
-  __ load_unsigned_byte(temp_reg, at_bcp(0));
-  __ cmpl(temp_reg, (int)Bytecodes::java_code(bc));
-  __ jccb(Assembler::equal, L_okay);
-  __ cmpl(temp_reg, bc_reg);
-  __ jcc(Assembler::equal, L_okay);
-  __ stop("patching the wrong bytecode");
-  __ bind(L_okay);
-#endif
-
-  // patch bytecode
-  __ movb(at_bcp(0), bc_reg);
-  __ bind(L_patch_done);
-}
-
-//----------------------------------------------------------------------------------------------------
-// Individual instructions
-
-void TemplateTable::nop() {
-  transition(vtos, vtos);
-  // nothing to do
-}
-
-void TemplateTable::shouldnotreachhere() {
-  transition(vtos, vtos);
-  __ stop("shouldnotreachhere bytecode");
-}
-
-
-
-void TemplateTable::aconst_null() {
-  transition(vtos, atos);
-  __ xorptr(rax, rax);
-}
-
-
-void TemplateTable::iconst(int value) {
-  transition(vtos, itos);
-  if (value == 0) {
-    __ xorptr(rax, rax);
-  } else {
-    __ movptr(rax, value);
-  }
-}
-
-
-void TemplateTable::lconst(int value) {
-  transition(vtos, ltos);
-  if (value == 0) {
-    __ xorptr(rax, rax);
-  } else {
-    __ movptr(rax, value);
-  }
-  assert(value >= 0, "check this code");
-  __ xorptr(rdx, rdx);
-}
-
-
-void TemplateTable::fconst(int value) {
-  transition(vtos, ftos);
-         if (value == 0) { __ fldz();
-  } else if (value == 1) { __ fld1();
-  } else if (value == 2) { __ fld1(); __ fld1(); __ faddp(); // should do a better solution here
-  } else                 { ShouldNotReachHere();
-  }
-}
-
-
-void TemplateTable::dconst(int value) {
-  transition(vtos, dtos);
-         if (value == 0) { __ fldz();
-  } else if (value == 1) { __ fld1();
-  } else                 { ShouldNotReachHere();
-  }
-}
-
-
-void TemplateTable::bipush() {
-  transition(vtos, itos);
-  __ load_signed_byte(rax, at_bcp(1));
-}
-
-
-void TemplateTable::sipush() {
-  transition(vtos, itos);
-  __ load_unsigned_short(rax, at_bcp(1));
-  __ bswapl(rax);
-  __ sarl(rax, 16);
-}
-
-void TemplateTable::ldc(bool wide) {
-  transition(vtos, vtos);
-  Label call_ldc, notFloat, notClass, Done;
-
-  if (wide) {
-    __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
-  } else {
-    __ load_unsigned_byte(rbx, at_bcp(1));
-  }
-  __ get_cpool_and_tags(rcx, rax);
-  const int base_offset = ConstantPool::header_size() * wordSize;
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-
-  // get type
-  __ xorptr(rdx, rdx);
-  __ movb(rdx, Address(rax, rbx, Address::times_1, tags_offset));
-
-  // unresolved class - get the resolved class
-  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClass);
-  __ jccb(Assembler::equal, call_ldc);
-
-  // unresolved class in error (resolution failed) - call into runtime
-  // so that the same error from first resolution attempt is thrown.
-  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClassInError);
-  __ jccb(Assembler::equal, call_ldc);
-
-  // resolved class - need to call vm to get java mirror of the class
-  __ cmpl(rdx, JVM_CONSTANT_Class);
-  __ jcc(Assembler::notEqual, notClass);
-
-  __ bind(call_ldc);
-  __ movl(rcx, wide);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::ldc), rcx);
-  __ push(atos);
-  __ jmp(Done);
-
-  __ bind(notClass);
-  __ cmpl(rdx, JVM_CONSTANT_Float);
-  __ jccb(Assembler::notEqual, notFloat);
-  // ftos
-  __ fld_s(    Address(rcx, rbx, Address::times_ptr, base_offset));
-  __ push(ftos);
-  __ jmp(Done);
-
-  __ bind(notFloat);
-#ifdef ASSERT
-  { Label L;
-    __ cmpl(rdx, JVM_CONSTANT_Integer);
-    __ jcc(Assembler::equal, L);
-    // String and Object are rewritten to fast_aldc
-    __ stop("unexpected tag type in ldc");
-    __ bind(L);
-  }
-#endif
-  // itos JVM_CONSTANT_Integer only
-  __ movl(rax, Address(rcx, rbx, Address::times_ptr, base_offset));
-  __ push(itos);
-  __ bind(Done);
-}
-
-// Fast path for caching oop constants.
-void TemplateTable::fast_aldc(bool wide) {
-  transition(vtos, atos);
-
-  Register result = rax;
-  Register tmp = rdx;
-  int index_size = wide ? sizeof(u2) : sizeof(u1);
-
-  Label resolved;
-
-  // We are resolved if the resolved reference cache entry contains a
-  // non-null object (String, MethodType, etc.)
-  assert_different_registers(result, tmp);
-  __ get_cache_index_at_bcp(tmp, 1, index_size);
-  __ load_resolved_reference_at_index(result, tmp);
-  __ testl(result, result);
-  __ jcc(Assembler::notZero, resolved);
-
-  address entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_ldc);
-
-  // first time invocation - must resolve first
-  __ movl(tmp, (int)bytecode());
-  __ call_VM(result, entry, tmp);
-
-  __ bind(resolved);
-
-  if (VerifyOops) {
-    __ verify_oop(result);
-  }
-}
-
-void TemplateTable::ldc2_w() {
-  transition(vtos, vtos);
-  Label Long, Done;
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
-
-  __ get_cpool_and_tags(rcx, rax);
-  const int base_offset = ConstantPool::header_size() * wordSize;
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-
-  // get type
-  __ cmpb(Address(rax, rbx, Address::times_1, tags_offset), JVM_CONSTANT_Double);
-  __ jccb(Assembler::notEqual, Long);
-  // dtos
-  __ fld_d(    Address(rcx, rbx, Address::times_ptr, base_offset));
-  __ push(dtos);
-  __ jmpb(Done);
-
-  __ bind(Long);
-  // ltos
-  __ movptr(rax, Address(rcx, rbx, Address::times_ptr, base_offset + 0 * wordSize));
-  NOT_LP64(__ movptr(rdx, Address(rcx, rbx, Address::times_ptr, base_offset + 1 * wordSize)));
-
-  __ push(ltos);
-
-  __ bind(Done);
-}
-
-
-void TemplateTable::locals_index(Register reg, int offset) {
-  __ load_unsigned_byte(reg, at_bcp(offset));
-  __ negptr(reg);
-}
-
-
-void TemplateTable::iload() {
-  transition(vtos, itos);
-  if (RewriteFrequentPairs) {
-    Label rewrite, done;
-
-    // get next byte
-    __ load_unsigned_byte(rbx, at_bcp(Bytecodes::length_for(Bytecodes::_iload)));
-    // if _iload, wait to rewrite to iload2.  We only want to rewrite the
-    // last two iloads in a pair.  Comparing against fast_iload means that
-    // the next bytecode is neither an iload or a caload, and therefore
-    // an iload pair.
-    __ cmpl(rbx, Bytecodes::_iload);
-    __ jcc(Assembler::equal, done);
-
-    __ cmpl(rbx, Bytecodes::_fast_iload);
-    __ movl(rcx, Bytecodes::_fast_iload2);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _caload, rewrite to fast_icaload
-    __ cmpl(rbx, Bytecodes::_caload);
-    __ movl(rcx, Bytecodes::_fast_icaload);
-    __ jccb(Assembler::equal, rewrite);
-
-    // rewrite so iload doesn't check again.
-    __ movl(rcx, Bytecodes::_fast_iload);
-
-    // rewrite
-    // rcx: fast bytecode
-    __ bind(rewrite);
-    patch_bytecode(Bytecodes::_iload, rcx, rbx, false);
-    __ bind(done);
-  }
-
-  // Get the local value into tos
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-
-void TemplateTable::fast_iload2() {
-  transition(vtos, itos);
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-  __ push(itos);
-  locals_index(rbx, 3);
-  __ movl(rax, iaddress(rbx));
-}
-
-void TemplateTable::fast_iload() {
-  transition(vtos, itos);
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-
-void TemplateTable::lload() {
-  transition(vtos, ltos);
-  locals_index(rbx);
-  __ movptr(rax, laddress(rbx));
-  NOT_LP64(__ movl(rdx, haddress(rbx)));
-}
-
-
-void TemplateTable::fload() {
-  transition(vtos, ftos);
-  locals_index(rbx);
-  __ fld_s(faddress(rbx));
-}
-
-
-void TemplateTable::dload() {
-  transition(vtos, dtos);
-  locals_index(rbx);
-  __ fld_d(daddress(rbx));
-}
-
-
-void TemplateTable::aload() {
-  transition(vtos, atos);
-  locals_index(rbx);
-  __ movptr(rax, aaddress(rbx));
-}
-
-
-void TemplateTable::locals_index_wide(Register reg) {
-  __ load_unsigned_short(reg, at_bcp(2));
-  __ bswapl(reg);
-  __ shrl(reg, 16);
-  __ negptr(reg);
-}
-
-
-void TemplateTable::wide_iload() {
-  transition(vtos, itos);
-  locals_index_wide(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-
-void TemplateTable::wide_lload() {
-  transition(vtos, ltos);
-  locals_index_wide(rbx);
-  __ movptr(rax, laddress(rbx));
-  NOT_LP64(__ movl(rdx, haddress(rbx)));
-}
-
-
-void TemplateTable::wide_fload() {
-  transition(vtos, ftos);
-  locals_index_wide(rbx);
-  __ fld_s(faddress(rbx));
-}
-
-
-void TemplateTable::wide_dload() {
-  transition(vtos, dtos);
-  locals_index_wide(rbx);
-  __ fld_d(daddress(rbx));
-}
-
-
-void TemplateTable::wide_aload() {
-  transition(vtos, atos);
-  locals_index_wide(rbx);
-  __ movptr(rax, aaddress(rbx));
-}
-
-void TemplateTable::index_check(Register array, Register index) {
-  // Pop ptr into array
-  __ pop_ptr(array);
-  index_check_without_pop(array, index);
-}
-
-void TemplateTable::index_check_without_pop(Register array, Register index) {
-  // destroys rbx,
-  // check array
-  __ null_check(array, arrayOopDesc::length_offset_in_bytes());
-  LP64_ONLY(__ movslq(index, index));
-  // check index
-  __ cmpl(index, Address(array, arrayOopDesc::length_offset_in_bytes()));
-  if (index != rbx) {
-    // ??? convention: move aberrant index into rbx, for exception message
-    assert(rbx != array, "different registers");
-    __ mov(rbx, index);
-  }
-  __ jump_cc(Assembler::aboveEqual,
-             ExternalAddress(Interpreter::_throw_ArrayIndexOutOfBoundsException_entry));
-}
-
-
-void TemplateTable::iaload() {
-  transition(itos, itos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  __ movl(rax, Address(rdx, rax, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_INT)));
-}
-
-
-void TemplateTable::laload() {
-  transition(itos, ltos);
-  // rax,: index
-  // rdx: array
-  index_check(rdx, rax);
-  __ mov(rbx, rax);
-  // rbx,: index
-  __ movptr(rax, Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 0 * wordSize));
-  NOT_LP64(__ movl(rdx, Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 1 * wordSize)));
-}
-
-
-void TemplateTable::faload() {
-  transition(itos, ftos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  __ fld_s(Address(rdx, rax, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_FLOAT)));
-}
-
-
-void TemplateTable::daload() {
-  transition(itos, dtos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  __ fld_d(Address(rdx, rax, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_DOUBLE)));
-}
-
-
-void TemplateTable::aaload() {
-  transition(itos, atos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  __ movptr(rax, Address(rdx, rax, Address::times_ptr, arrayOopDesc::base_offset_in_bytes(T_OBJECT)));
-}
-
-
-void TemplateTable::baload() {
-  transition(itos, itos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  // can do better code for P5 - fix this at some point
-  __ load_signed_byte(rbx, Address(rdx, rax, Address::times_1, arrayOopDesc::base_offset_in_bytes(T_BYTE)));
-  __ mov(rax, rbx);
-}
-
-
-void TemplateTable::caload() {
-  transition(itos, itos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  // can do better code for P5 - may want to improve this at some point
-  __ load_unsigned_short(rbx, Address(rdx, rax, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_CHAR)));
-  __ mov(rax, rbx);
-}
-
-// iload followed by caload frequent pair
-void TemplateTable::fast_icaload() {
-  transition(vtos, itos);
-  // load index out of locals
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-
-  // rdx: array
-  index_check(rdx, rax);
-  // rax,: index
-  __ load_unsigned_short(rbx, Address(rdx, rax, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_CHAR)));
-  __ mov(rax, rbx);
-}
-
-void TemplateTable::saload() {
-  transition(itos, itos);
-  // rdx: array
-  index_check(rdx, rax);  // kills rbx,
-  // rax,: index
-  // can do better code for P5 - may want to improve this at some point
-  __ load_signed_short(rbx, Address(rdx, rax, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_SHORT)));
-  __ mov(rax, rbx);
-}
-
-
-void TemplateTable::iload(int n) {
-  transition(vtos, itos);
-  __ movl(rax, iaddress(n));
-}
-
-
-void TemplateTable::lload(int n) {
-  transition(vtos, ltos);
-  __ movptr(rax, laddress(n));
-  NOT_LP64(__ movptr(rdx, haddress(n)));
-}
-
-
-void TemplateTable::fload(int n) {
-  transition(vtos, ftos);
-  __ fld_s(faddress(n));
-}
-
-
-void TemplateTable::dload(int n) {
-  transition(vtos, dtos);
-  __ fld_d(daddress(n));
-}
-
-
-void TemplateTable::aload(int n) {
-  transition(vtos, atos);
-  __ movptr(rax, aaddress(n));
-}
-
-
-void TemplateTable::aload_0() {
-  transition(vtos, atos);
-  // According to bytecode histograms, the pairs:
-  //
-  // _aload_0, _fast_igetfield
-  // _aload_0, _fast_agetfield
-  // _aload_0, _fast_fgetfield
-  //
-  // occur frequently. If RewriteFrequentPairs is set, the (slow) _aload_0
-  // bytecode checks if the next bytecode is either _fast_igetfield,
-  // _fast_agetfield or _fast_fgetfield and then rewrites the
-  // current bytecode into a pair bytecode; otherwise it rewrites the current
-  // bytecode into _fast_aload_0 that doesn't do the pair check anymore.
-  //
-  // Note: If the next bytecode is _getfield, the rewrite must be delayed,
-  //       otherwise we may miss an opportunity for a pair.
-  //
-  // Also rewrite frequent pairs
-  //   aload_0, aload_1
-  //   aload_0, iload_1
-  // These bytecodes with a small amount of code are most profitable to rewrite
-  if (RewriteFrequentPairs) {
-    Label rewrite, done;
-    // get next byte
-    __ load_unsigned_byte(rbx, at_bcp(Bytecodes::length_for(Bytecodes::_aload_0)));
-
-    // do actual aload_0
-    aload(0);
-
-    // if _getfield then wait with rewrite
-    __ cmpl(rbx, Bytecodes::_getfield);
-    __ jcc(Assembler::equal, done);
-
-    // if _igetfield then reqrite to _fast_iaccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_iaccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_igetfield);
-    __ movl(rcx, Bytecodes::_fast_iaccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _agetfield then reqrite to _fast_aaccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_aaccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_agetfield);
-    __ movl(rcx, Bytecodes::_fast_aaccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _fgetfield then reqrite to _fast_faccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_faccess_0) == Bytecodes::_aload_0, "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_fgetfield);
-    __ movl(rcx, Bytecodes::_fast_faccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // else rewrite to _fast_aload0
-    assert(Bytecodes::java_code(Bytecodes::_fast_aload_0) == Bytecodes::_aload_0, "fix bytecode definition");
-    __ movl(rcx, Bytecodes::_fast_aload_0);
-
-    // rewrite
-    // rcx: fast bytecode
-    __ bind(rewrite);
-    patch_bytecode(Bytecodes::_aload_0, rcx, rbx, false);
-
-    __ bind(done);
-  } else {
-    aload(0);
-  }
-}
-
-void TemplateTable::istore() {
-  transition(itos, vtos);
-  locals_index(rbx);
-  __ movl(iaddress(rbx), rax);
-}
-
-
-void TemplateTable::lstore() {
-  transition(ltos, vtos);
-  locals_index(rbx);
-  __ movptr(laddress(rbx), rax);
-  NOT_LP64(__ movptr(haddress(rbx), rdx));
-}
-
-
-void TemplateTable::fstore() {
-  transition(ftos, vtos);
-  locals_index(rbx);
-  __ fstp_s(faddress(rbx));
-}
-
-
-void TemplateTable::dstore() {
-  transition(dtos, vtos);
-  locals_index(rbx);
-  __ fstp_d(daddress(rbx));
-}
-
-
-void TemplateTable::astore() {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  locals_index(rbx);
-  __ movptr(aaddress(rbx), rax);
-}
-
-
-void TemplateTable::wide_istore() {
-  transition(vtos, vtos);
-  __ pop_i(rax);
-  locals_index_wide(rbx);
-  __ movl(iaddress(rbx), rax);
-}
-
-
-void TemplateTable::wide_lstore() {
-  transition(vtos, vtos);
-  __ pop_l(rax, rdx);
-  locals_index_wide(rbx);
-  __ movptr(laddress(rbx), rax);
-  NOT_LP64(__ movl(haddress(rbx), rdx));
-}
-
-
-void TemplateTable::wide_fstore() {
-  wide_istore();
-}
-
-
-void TemplateTable::wide_dstore() {
-  wide_lstore();
-}
-
-
-void TemplateTable::wide_astore() {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  locals_index_wide(rbx);
-  __ movptr(aaddress(rbx), rax);
-}
-
-
-void TemplateTable::iastore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  // rax,: value
-  // rdx: array
-  index_check(rdx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ movl(Address(rdx, rbx, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_INT)), rax);
-}
-
-
-void TemplateTable::lastore() {
-  transition(ltos, vtos);
-  __ pop_i(rbx);
-  // rax,: low(value)
-  // rcx: array
-  // rdx: high(value)
-  index_check(rcx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ movptr(Address(rcx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 0 * wordSize), rax);
-  NOT_LP64(__ movl(Address(rcx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_LONG) + 1 * wordSize), rdx));
-}
-
-
-void TemplateTable::fastore() {
-  transition(ftos, vtos);
-  __ pop_i(rbx);
-  // rdx: array
-  // st0: value
-  index_check(rdx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ fstp_s(Address(rdx, rbx, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_FLOAT)));
-}
-
-
-void TemplateTable::dastore() {
-  transition(dtos, vtos);
-  __ pop_i(rbx);
-  // rdx: array
-  // st0: value
-  index_check(rdx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ fstp_d(Address(rdx, rbx, Address::times_8, arrayOopDesc::base_offset_in_bytes(T_DOUBLE)));
-}
-
-
-void TemplateTable::aastore() {
-  Label is_null, ok_is_subtype, done;
-  transition(vtos, vtos);
-  // stack: ..., array, index, value
-  __ movptr(rax, at_tos());     // Value
-  __ movl(rcx, at_tos_p1());  // Index
-  __ movptr(rdx, at_tos_p2());  // Array
-
-  Address element_address(rdx, rcx, Address::times_4, arrayOopDesc::base_offset_in_bytes(T_OBJECT));
-  index_check_without_pop(rdx, rcx);      // kills rbx,
-  // do array store check - check for NULL value first
-  __ testptr(rax, rax);
-  __ jcc(Assembler::zero, is_null);
-
-  // Move subklass into EBX
-  __ load_klass(rbx, rax);
-  // Move superklass into EAX
-  __ load_klass(rax, rdx);
-  __ movptr(rax, Address(rax, ObjArrayKlass::element_klass_offset()));
-  // Compress array+index*wordSize+12 into a single register.  Frees ECX.
-  __ lea(rdx, element_address);
-
-  // Generate subtype check.  Blows ECX.  Resets EDI to locals.
-  // Superklass in EAX.  Subklass in EBX.
-  __ gen_subtype_check( rbx, ok_is_subtype );
-
-  // Come here on failure
-  // object is at TOS
-  __ jump(ExternalAddress(Interpreter::_throw_ArrayStoreException_entry));
-
-  // Come here on success
-  __ bind(ok_is_subtype);
-
-  // Get the value to store
-  __ movptr(rax, at_rsp());
-  // and store it with appropriate barrier
-  do_oop_store(_masm, Address(rdx, 0), rax, _bs->kind(), true);
-
-  __ jmp(done);
-
-  // Have a NULL in EAX, EDX=array, ECX=index.  Store NULL at ary[idx]
-  __ bind(is_null);
-  __ profile_null_seen(rbx);
-
-  // Store NULL, (noreg means NULL to do_oop_store)
-  do_oop_store(_masm, element_address, noreg, _bs->kind(), true);
-
-  // Pop stack arguments
-  __ bind(done);
-  __ addptr(rsp, 3 * Interpreter::stackElementSize);
-}
-
-
-void TemplateTable::bastore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  // rax,: value
-  // rdx: array
-  index_check(rdx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ movb(Address(rdx, rbx, Address::times_1, arrayOopDesc::base_offset_in_bytes(T_BYTE)), rax);
-}
-
-
-void TemplateTable::castore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  // rax,: value
-  // rdx: array
-  index_check(rdx, rbx);  // prefer index in rbx,
-  // rbx,: index
-  __ movw(Address(rdx, rbx, Address::times_2, arrayOopDesc::base_offset_in_bytes(T_CHAR)), rax);
-}
-
-
-void TemplateTable::sastore() {
-  castore();
-}
-
-
-void TemplateTable::istore(int n) {
-  transition(itos, vtos);
-  __ movl(iaddress(n), rax);
-}
-
-
-void TemplateTable::lstore(int n) {
-  transition(ltos, vtos);
-  __ movptr(laddress(n), rax);
-  NOT_LP64(__ movptr(haddress(n), rdx));
-}
-
-
-void TemplateTable::fstore(int n) {
-  transition(ftos, vtos);
-  __ fstp_s(faddress(n));
-}
-
-
-void TemplateTable::dstore(int n) {
-  transition(dtos, vtos);
-  __ fstp_d(daddress(n));
-}
-
-
-void TemplateTable::astore(int n) {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  __ movptr(aaddress(n), rax);
-}
-
-
-void TemplateTable::pop() {
-  transition(vtos, vtos);
-  __ addptr(rsp, Interpreter::stackElementSize);
-}
-
-
-void TemplateTable::pop2() {
-  transition(vtos, vtos);
-  __ addptr(rsp, 2*Interpreter::stackElementSize);
-}
-
-
-void TemplateTable::dup() {
-  transition(vtos, vtos);
-  // stack: ..., a
-  __ load_ptr(0, rax);
-  __ push_ptr(rax);
-  // stack: ..., a, a
-}
-
-
-void TemplateTable::dup_x1() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr( 0, rax);  // load b
-  __ load_ptr( 1, rcx);  // load a
-  __ store_ptr(1, rax);  // store b
-  __ store_ptr(0, rcx);  // store a
-  __ push_ptr(rax);      // push b
-  // stack: ..., b, a, b
-}
-
-
-void TemplateTable::dup_x2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c
-  __ load_ptr( 0, rax);  // load c
-  __ load_ptr( 2, rcx);  // load a
-  __ store_ptr(2, rax);  // store c in a
-  __ push_ptr(rax);      // push c
-  // stack: ..., c, b, c, c
-  __ load_ptr( 2, rax);  // load b
-  __ store_ptr(2, rcx);  // store a in b
-  // stack: ..., c, a, c, c
-  __ store_ptr(1, rax);  // store b in c
-  // stack: ..., c, a, b, c
-}
-
-
-void TemplateTable::dup2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr(1, rax);  // load a
-  __ push_ptr(rax);     // push a
-  __ load_ptr(1, rax);  // load b
-  __ push_ptr(rax);     // push b
-  // stack: ..., a, b, a, b
-}
-
-
-void TemplateTable::dup2_x1() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c
-  __ load_ptr( 0, rcx);  // load c
-  __ load_ptr( 1, rax);  // load b
-  __ push_ptr(rax);      // push b
-  __ push_ptr(rcx);      // push c
-  // stack: ..., a, b, c, b, c
-  __ store_ptr(3, rcx);  // store c in b
-  // stack: ..., a, c, c, b, c
-  __ load_ptr( 4, rcx);  // load a
-  __ store_ptr(2, rcx);  // store a in 2nd c
-  // stack: ..., a, c, a, b, c
-  __ store_ptr(4, rax);  // store b in a
-  // stack: ..., b, c, a, b, c
-  // stack: ..., b, c, a, b, c
-}
-
-
-void TemplateTable::dup2_x2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c, d
-  __ load_ptr( 0, rcx);  // load d
-  __ load_ptr( 1, rax);  // load c
-  __ push_ptr(rax);      // push c
-  __ push_ptr(rcx);      // push d
-  // stack: ..., a, b, c, d, c, d
-  __ load_ptr( 4, rax);  // load b
-  __ store_ptr(2, rax);  // store b in d
-  __ store_ptr(4, rcx);  // store d in b
-  // stack: ..., a, d, c, b, c, d
-  __ load_ptr( 5, rcx);  // load a
-  __ load_ptr( 3, rax);  // load c
-  __ store_ptr(3, rcx);  // store a in c
-  __ store_ptr(5, rax);  // store c in a
-  // stack: ..., c, d, a, b, c, d
-  // stack: ..., c, d, a, b, c, d
-}
-
-
-void TemplateTable::swap() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr( 1, rcx);  // load a
-  __ load_ptr( 0, rax);  // load b
-  __ store_ptr(0, rcx);  // store a in b
-  __ store_ptr(1, rax);  // store b in a
-  // stack: ..., b, a
-}
-
-
-void TemplateTable::iop2(Operation op) {
-  transition(itos, itos);
-  switch (op) {
-    case add  :                   __ pop_i(rdx); __ addl (rax, rdx); break;
-    case sub  : __ mov(rdx, rax); __ pop_i(rax); __ subl (rax, rdx); break;
-    case mul  :                   __ pop_i(rdx); __ imull(rax, rdx); break;
-    case _and :                   __ pop_i(rdx); __ andl (rax, rdx); break;
-    case _or  :                   __ pop_i(rdx); __ orl  (rax, rdx); break;
-    case _xor :                   __ pop_i(rdx); __ xorl (rax, rdx); break;
-    case shl  : __ mov(rcx, rax); __ pop_i(rax); __ shll (rax);      break; // implicit masking of lower 5 bits by Intel shift instr.
-    case shr  : __ mov(rcx, rax); __ pop_i(rax); __ sarl (rax);      break; // implicit masking of lower 5 bits by Intel shift instr.
-    case ushr : __ mov(rcx, rax); __ pop_i(rax); __ shrl (rax);      break; // implicit masking of lower 5 bits by Intel shift instr.
-    default   : ShouldNotReachHere();
-  }
-}
-
-
-void TemplateTable::lop2(Operation op) {
-  transition(ltos, ltos);
-  __ pop_l(rbx, rcx);
-  switch (op) {
-    case add  : __ addl(rax, rbx); __ adcl(rdx, rcx); break;
-    case sub  : __ subl(rbx, rax); __ sbbl(rcx, rdx);
-                __ mov (rax, rbx); __ mov (rdx, rcx); break;
-    case _and : __ andl(rax, rbx); __ andl(rdx, rcx); break;
-    case _or  : __ orl (rax, rbx); __ orl (rdx, rcx); break;
-    case _xor : __ xorl(rax, rbx); __ xorl(rdx, rcx); break;
-    default   : ShouldNotReachHere();
-  }
-}
-
-
-void TemplateTable::idiv() {
-  transition(itos, itos);
-  __ mov(rcx, rax);
-  __ pop_i(rax);
-  // Note: could xor rax, and rcx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivl(rcx);
-}
-
-
-void TemplateTable::irem() {
-  transition(itos, itos);
-  __ mov(rcx, rax);
-  __ pop_i(rax);
-  // Note: could xor rax, and rcx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivl(rcx);
-  __ mov(rax, rdx);
-}
-
-
-void TemplateTable::lmul() {
-  transition(ltos, ltos);
-  __ pop_l(rbx, rcx);
-  __ push(rcx); __ push(rbx);
-  __ push(rdx); __ push(rax);
-  __ lmul(2 * wordSize, 0);
-  __ addptr(rsp, 4 * wordSize);  // take off temporaries
-}
-
-
-void TemplateTable::ldiv() {
-  transition(ltos, ltos);
-  __ pop_l(rbx, rcx);
-  __ push(rcx); __ push(rbx);
-  __ push(rdx); __ push(rax);
-  // check if y = 0
-  __ orl(rax, rdx);
-  __ jump_cc(Assembler::zero,
-             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
-  __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::ldiv));
-  __ addptr(rsp, 4 * wordSize);  // take off temporaries
-}
-
-
-void TemplateTable::lrem() {
-  transition(ltos, ltos);
-  __ pop_l(rbx, rcx);
-  __ push(rcx); __ push(rbx);
-  __ push(rdx); __ push(rax);
-  // check if y = 0
-  __ orl(rax, rdx);
-  __ jump_cc(Assembler::zero,
-             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
-  __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::lrem));
-  __ addptr(rsp, 4 * wordSize);
-}
-
-
-void TemplateTable::lshl() {
-  transition(itos, ltos);
-  __ movl(rcx, rax);                             // get shift count
-  __ pop_l(rax, rdx);                            // get shift value
-  __ lshl(rdx, rax);
-}
-
-
-void TemplateTable::lshr() {
-  transition(itos, ltos);
-  __ mov(rcx, rax);                              // get shift count
-  __ pop_l(rax, rdx);                            // get shift value
-  __ lshr(rdx, rax, true);
-}
-
-
-void TemplateTable::lushr() {
-  transition(itos, ltos);
-  __ mov(rcx, rax);                              // get shift count
-  __ pop_l(rax, rdx);                            // get shift value
-  __ lshr(rdx, rax);
-}
-
-
-void TemplateTable::fop2(Operation op) {
-  transition(ftos, ftos);
-  switch (op) {
-    case add: __ fadd_s (at_rsp());                break;
-    case sub: __ fsubr_s(at_rsp());                break;
-    case mul: __ fmul_s (at_rsp());                break;
-    case div: __ fdivr_s(at_rsp());                break;
-    case rem: __ fld_s  (at_rsp()); __ fremr(rax); break;
-    default : ShouldNotReachHere();
-  }
-  __ f2ieee();
-  __ pop(rax);  // pop float thing off
-}
-
-
-void TemplateTable::dop2(Operation op) {
-  transition(dtos, dtos);
-
-  switch (op) {
-    case add: __ fadd_d (at_rsp());                break;
-    case sub: __ fsubr_d(at_rsp());                break;
-    case mul: {
-      Label L_strict;
-      Label L_join;
-      const Address access_flags      (rcx, Method::access_flags_offset());
-      __ get_method(rcx);
-      __ movl(rcx, access_flags);
-      __ testl(rcx, JVM_ACC_STRICT);
-      __ jccb(Assembler::notZero, L_strict);
-      __ fmul_d (at_rsp());
-      __ jmpb(L_join);
-      __ bind(L_strict);
-      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias1()));
-      __ fmulp();
-      __ fmul_d (at_rsp());
-      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias2()));
-      __ fmulp();
-      __ bind(L_join);
-      break;
-    }
-    case div: {
-      Label L_strict;
-      Label L_join;
-      const Address access_flags      (rcx, Method::access_flags_offset());
-      __ get_method(rcx);
-      __ movl(rcx, access_flags);
-      __ testl(rcx, JVM_ACC_STRICT);
-      __ jccb(Assembler::notZero, L_strict);
-      __ fdivr_d(at_rsp());
-      __ jmp(L_join);
-      __ bind(L_strict);
-      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias1()));
-      __ fmul_d (at_rsp());
-      __ fdivrp();
-      __ fld_x(ExternalAddress(StubRoutines::addr_fpu_subnormal_bias2()));
-      __ fmulp();
-      __ bind(L_join);
-      break;
-    }
-    case rem: __ fld_d  (at_rsp()); __ fremr(rax); break;
-    default : ShouldNotReachHere();
-  }
-  __ d2ieee();
-  // Pop double precision number from rsp.
-  __ pop(rax);
-  __ pop(rdx);
-}
-
-
-void TemplateTable::ineg() {
-  transition(itos, itos);
-  __ negl(rax);
-}
-
-
-void TemplateTable::lneg() {
-  transition(ltos, ltos);
-  __ lneg(rdx, rax);
-}
-
-
-void TemplateTable::fneg() {
-  transition(ftos, ftos);
-  __ fchs();
-}
-
-
-void TemplateTable::dneg() {
-  transition(dtos, dtos);
-  __ fchs();
-}
-
-
-void TemplateTable::iinc() {
-  transition(vtos, vtos);
-  __ load_signed_byte(rdx, at_bcp(2));           // get constant
-  locals_index(rbx);
-  __ addl(iaddress(rbx), rdx);
-}
-
-
-void TemplateTable::wide_iinc() {
-  transition(vtos, vtos);
-  __ movl(rdx, at_bcp(4));                       // get constant
-  locals_index_wide(rbx);
-  __ bswapl(rdx);                                 // swap bytes & sign-extend constant
-  __ sarl(rdx, 16);
-  __ addl(iaddress(rbx), rdx);
-  // Note: should probably use only one movl to get both
-  //       the index and the constant -> fix this
-}
-
-
-void TemplateTable::convert() {
-  // Checking
-#ifdef ASSERT
-  { TosState tos_in  = ilgl;
-    TosState tos_out = ilgl;
-    switch (bytecode()) {
-      case Bytecodes::_i2l: // fall through
-      case Bytecodes::_i2f: // fall through
-      case Bytecodes::_i2d: // fall through
-      case Bytecodes::_i2b: // fall through
-      case Bytecodes::_i2c: // fall through
-      case Bytecodes::_i2s: tos_in = itos; break;
-      case Bytecodes::_l2i: // fall through
-      case Bytecodes::_l2f: // fall through
-      case Bytecodes::_l2d: tos_in = ltos; break;
-      case Bytecodes::_f2i: // fall through
-      case Bytecodes::_f2l: // fall through
-      case Bytecodes::_f2d: tos_in = ftos; break;
-      case Bytecodes::_d2i: // fall through
-      case Bytecodes::_d2l: // fall through
-      case Bytecodes::_d2f: tos_in = dtos; break;
-      default             : ShouldNotReachHere();
-    }
-    switch (bytecode()) {
-      case Bytecodes::_l2i: // fall through
-      case Bytecodes::_f2i: // fall through
-      case Bytecodes::_d2i: // fall through
-      case Bytecodes::_i2b: // fall through
-      case Bytecodes::_i2c: // fall through
-      case Bytecodes::_i2s: tos_out = itos; break;
-      case Bytecodes::_i2l: // fall through
-      case Bytecodes::_f2l: // fall through
-      case Bytecodes::_d2l: tos_out = ltos; break;
-      case Bytecodes::_i2f: // fall through
-      case Bytecodes::_l2f: // fall through
-      case Bytecodes::_d2f: tos_out = ftos; break;
-      case Bytecodes::_i2d: // fall through
-      case Bytecodes::_l2d: // fall through
-      case Bytecodes::_f2d: tos_out = dtos; break;
-      default             : ShouldNotReachHere();
-    }
-    transition(tos_in, tos_out);
-  }
-#endif // ASSERT
-
-  // Conversion
-  // (Note: use push(rcx)/pop(rcx) for 1/2-word stack-ptr manipulation)
-  switch (bytecode()) {
-    case Bytecodes::_i2l:
-      __ extend_sign(rdx, rax);
-      break;
-    case Bytecodes::_i2f:
-      __ push(rax);          // store int on tos
-      __ fild_s(at_rsp());   // load int to ST0
-      __ f2ieee();           // truncate to float size
-      __ pop(rcx);           // adjust rsp
-      break;
-    case Bytecodes::_i2d:
-      __ push(rax);          // add one slot for d2ieee()
-      __ push(rax);          // store int on tos
-      __ fild_s(at_rsp());   // load int to ST0
-      __ d2ieee();           // truncate to double size
-      __ pop(rcx);           // adjust rsp
-      __ pop(rcx);
-      break;
-    case Bytecodes::_i2b:
-      __ shll(rax, 24);      // truncate upper 24 bits
-      __ sarl(rax, 24);      // and sign-extend byte
-      LP64_ONLY(__ movsbl(rax, rax));
-      break;
-    case Bytecodes::_i2c:
-      __ andl(rax, 0xFFFF);  // truncate upper 16 bits
-      LP64_ONLY(__ movzwl(rax, rax));
-      break;
-    case Bytecodes::_i2s:
-      __ shll(rax, 16);      // truncate upper 16 bits
-      __ sarl(rax, 16);      // and sign-extend short
-      LP64_ONLY(__ movswl(rax, rax));
-      break;
-    case Bytecodes::_l2i:
-      /* nothing to do */
-      break;
-    case Bytecodes::_l2f:
-      __ push(rdx);          // store long on tos
-      __ push(rax);
-      __ fild_d(at_rsp());   // load long to ST0
-      __ f2ieee();           // truncate to float size
-      __ pop(rcx);           // adjust rsp
-      __ pop(rcx);
-      break;
-    case Bytecodes::_l2d:
-      __ push(rdx);          // store long on tos
-      __ push(rax);
-      __ fild_d(at_rsp());   // load long to ST0
-      __ d2ieee();           // truncate to double size
-      __ pop(rcx);           // adjust rsp
-      __ pop(rcx);
-      break;
-    case Bytecodes::_f2i:
-      __ push(rcx);          // reserve space for argument
-      __ fstp_s(at_rsp());   // pass float argument on stack
-      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2i), 1);
-      break;
-    case Bytecodes::_f2l:
-      __ push(rcx);          // reserve space for argument
-      __ fstp_s(at_rsp());   // pass float argument on stack
-      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2l), 1);
-      break;
-    case Bytecodes::_f2d:
-      /* nothing to do */
-      break;
-    case Bytecodes::_d2i:
-      __ push(rcx);          // reserve space for argument
-      __ push(rcx);
-      __ fstp_d(at_rsp());   // pass double argument on stack
-      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2i), 2);
-      break;
-    case Bytecodes::_d2l:
-      __ push(rcx);          // reserve space for argument
-      __ push(rcx);
-      __ fstp_d(at_rsp());   // pass double argument on stack
-      __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2l), 2);
-      break;
-    case Bytecodes::_d2f:
-      __ push(rcx);          // reserve space for f2ieee()
-      __ f2ieee();           // truncate to float size
-      __ pop(rcx);           // adjust rsp
-      break;
-    default             :
-      ShouldNotReachHere();
-  }
-}
-
-
-void TemplateTable::lcmp() {
-  transition(ltos, itos);
-  // y = rdx:rax
-  __ pop_l(rbx, rcx);             // get x = rcx:rbx
-  __ lcmp2int(rcx, rbx, rdx, rax);// rcx := cmp(x, y)
-  __ mov(rax, rcx);
-}
-
-
-void TemplateTable::float_cmp(bool is_float, int unordered_result) {
-  if (is_float) {
-    __ fld_s(at_rsp());
-  } else {
-    __ fld_d(at_rsp());
-    __ pop(rdx);
-  }
-  __ pop(rcx);
-  __ fcmp2int(rax, unordered_result < 0);
-}
-
-
-void TemplateTable::branch(bool is_jsr, bool is_wide) {
-  __ get_method(rcx);           // ECX holds method
-  __ profile_taken_branch(rax,rbx); // EAX holds updated MDP, EBX holds bumped taken count
-
-  const ByteSize be_offset = MethodCounters::backedge_counter_offset() +
-                             InvocationCounter::counter_offset();
-  const ByteSize inv_offset = MethodCounters::invocation_counter_offset() +
-                              InvocationCounter::counter_offset();
-
-  // Load up EDX with the branch displacement
-  if (is_wide) {
-    __ movl(rdx, at_bcp(1));
-  } else {
-    __ load_signed_short(rdx, at_bcp(1));
-  }
-  __ bswapl(rdx);
-  if (!is_wide) __ sarl(rdx, 16);
-  LP64_ONLY(__ movslq(rdx, rdx));
-
-
-  // Handle all the JSR stuff here, then exit.
-  // It's much shorter and cleaner than intermingling with the
-  // non-JSR normal-branch stuff occurring below.
-  if (is_jsr) {
-    // Pre-load the next target bytecode into EBX
-    __ load_unsigned_byte(rbx, Address(rsi, rdx, Address::times_1, 0));
-
-    // compute return address as bci in rax,
-    __ lea(rax, at_bcp((is_wide ? 5 : 3) - in_bytes(ConstMethod::codes_offset())));
-    __ subptr(rax, Address(rcx, Method::const_offset()));
-    // Adjust the bcp in RSI by the displacement in EDX
-    __ addptr(rsi, rdx);
-    // Push return address
-    __ push_i(rax);
-    // jsr returns vtos
-    __ dispatch_only_noverify(vtos);
-    return;
-  }
-
-  // Normal (non-jsr) branch handling
-
-  // Adjust the bcp in RSI by the displacement in EDX
-  __ addptr(rsi, rdx);
-
-  assert(UseLoopCounter || !UseOnStackReplacement, "on-stack-replacement requires loop counters");
-  Label backedge_counter_overflow;
-  Label profile_method;
-  Label dispatch;
-  if (UseLoopCounter) {
-    // increment backedge counter for backward branches
-    // rax,: MDO
-    // rbx,: MDO bumped taken-count
-    // rcx: method
-    // rdx: target offset
-    // rsi: target bcp
-    // rdi: locals pointer
-    __ testl(rdx, rdx);             // check if forward or backward branch
-    __ jcc(Assembler::positive, dispatch); // count only if backward branch
-
-    // check if MethodCounters exists
-    Label has_counters;
-    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
-    __ testptr(rax, rax);
-    __ jcc(Assembler::notZero, has_counters);
-    __ push(rdx);
-    __ push(rcx);
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::build_method_counters),
-               rcx);
-    __ pop(rcx);
-    __ pop(rdx);
-    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
-    __ testptr(rax, rax);
-    __ jcc(Assembler::zero, dispatch);
-    __ bind(has_counters);
-
-    if (TieredCompilation) {
-      Label no_mdo;
-      int increment = InvocationCounter::count_increment;
-      if (ProfileInterpreter) {
-        // Are we profiling?
-        __ movptr(rbx, Address(rcx, in_bytes(Method::method_data_offset())));
-        __ testptr(rbx, rbx);
-        __ jccb(Assembler::zero, no_mdo);
-        // Increment the MDO backedge counter
-        const Address mdo_backedge_counter(rbx, in_bytes(MethodData::backedge_counter_offset()) +
-                                                in_bytes(InvocationCounter::counter_offset()));
-        const Address mask(rbx, in_bytes(MethodData::backedge_mask_offset()));
-        __ increment_mask_and_jump(mdo_backedge_counter, increment, mask,
-                                   rax, false, Assembler::zero, &backedge_counter_overflow);
-        __ jmp(dispatch);
-      }
-      __ bind(no_mdo);
-      // Increment backedge counter in MethodCounters*
-      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
-      const Address mask(rcx, in_bytes(MethodCounters::backedge_mask_offset()));
-      __ increment_mask_and_jump(Address(rcx, be_offset), increment, mask,
-                                 rax, false, Assembler::zero, &backedge_counter_overflow);
-    } else { // not TieredCompilation
-      // increment counter
-      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
-      __ movl(rax, Address(rcx, be_offset));        // load backedge counter
-      __ incrementl(rax, InvocationCounter::count_increment); // increment counter
-      __ movl(Address(rcx, be_offset), rax);        // store counter
-
-      __ movl(rax, Address(rcx, inv_offset));    // load invocation counter
-
-      __ andl(rax, InvocationCounter::count_mask_value);     // and the status bits
-      __ addl(rax, Address(rcx, be_offset));        // add both counters
-
-      if (ProfileInterpreter) {
-        // Test to see if we should create a method data oop
-        __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_profile_limit_offset())));
-        __ jcc(Assembler::less, dispatch);
-
-        // if no method data exists, go to profile method
-        __ test_method_data_pointer(rax, profile_method);
-
-        if (UseOnStackReplacement) {
-          // check for overflow against rbx, which is the MDO taken count
-          __ cmp32(rbx, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
-          __ jcc(Assembler::below, dispatch);
-
-          // When ProfileInterpreter is on, the backedge_count comes from the
-          // MethodData*, which value does not get reset on the call to
-          // frequency_counter_overflow().  To avoid excessive calls to the overflow
-          // routine while the method is being compiled, add a second test to make
-          // sure the overflow function is called only once every overflow_frequency.
-          const int overflow_frequency = 1024;
-          __ andptr(rbx, overflow_frequency-1);
-          __ jcc(Assembler::zero, backedge_counter_overflow);
-        }
-      } else {
-        if (UseOnStackReplacement) {
-          // check for overflow against rax, which is the sum of the counters
-          __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
-          __ jcc(Assembler::aboveEqual, backedge_counter_overflow);
-
-        }
-      }
-    }
-    __ bind(dispatch);
-  }
-
-  // Pre-load the next target bytecode into EBX
-  __ load_unsigned_byte(rbx, Address(rsi, 0));
-
-  // continue with the bytecode @ target
-  // rax,: return bci for jsr's, unused otherwise
-  // rbx,: target bytecode
-  // rsi: target bcp
-  __ dispatch_only(vtos);
-
-  if (UseLoopCounter) {
-    if (ProfileInterpreter) {
-      // Out-of-line code to allocate method data oop.
-      __ bind(profile_method);
-      __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::profile_method));
-      __ load_unsigned_byte(rbx, Address(rsi, 0));  // restore target bytecode
-      __ set_method_data_pointer_for_bcp();
-      __ jmp(dispatch);
-    }
-
-    if (UseOnStackReplacement) {
-
-      // invocation counter overflow
-      __ bind(backedge_counter_overflow);
-      __ negptr(rdx);
-      __ addptr(rdx, rsi);        // branch bcp
-      call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::frequency_counter_overflow), rdx);
-      __ load_unsigned_byte(rbx, Address(rsi, 0));  // restore target bytecode
-
-      // rax,: osr nmethod (osr ok) or NULL (osr not possible)
-      // rbx,: target bytecode
-      // rdx: scratch
-      // rdi: locals pointer
-      // rsi: bcp
-      __ testptr(rax, rax);                      // test result
-      __ jcc(Assembler::zero, dispatch);         // no osr if null
-      // nmethod may have been invalidated (VM may block upon call_VM return)
-      __ cmpb(Address(rax, nmethod::state_offset()), nmethod::in_use);
-      __ jcc(Assembler::notEqual, dispatch);
-
-      // We have the address of an on stack replacement routine in rax,
-      // We need to prepare to execute the OSR method. First we must
-      // migrate the locals and monitors off of the stack.
-
-      __ mov(rbx, rax);                             // save the nmethod
-
-      __ get_thread(rcx);
-      call_VM(noreg, CAST_FROM_FN_PTR(address, SharedRuntime::OSR_migration_begin));
-      // rax, is OSR buffer, move it to expected parameter location
-      __ mov(rcx, rax);
-
-      // pop the interpreter frame
-      __ movptr(rdx, Address(rbp, frame::interpreter_frame_sender_sp_offset * wordSize)); // get sender sp
-      __ leave();                                // remove frame anchor
-      __ pop(rdi);                               // get return address
-      __ mov(rsp, rdx);                          // set sp to sender sp
-
-      // Align stack pointer for compiled code (note that caller is
-      // responsible for undoing this fixup by remembering the old SP
-      // in an rbp,-relative location)
-      __ andptr(rsp, -(StackAlignmentInBytes));
-
-      // push the (possibly adjusted) return address
-      __ push(rdi);
-
-      // and begin the OSR nmethod
-      __ jmp(Address(rbx, nmethod::osr_entry_point_offset()));
-    }
-  }
-}
-
-
-void TemplateTable::if_0cmp(Condition cc) {
-  transition(itos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ testl(rax, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-
-void TemplateTable::if_icmp(Condition cc) {
-  transition(itos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ pop_i(rdx);
-  __ cmpl(rdx, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-
-void TemplateTable::if_nullcmp(Condition cc) {
-  transition(atos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ testptr(rax, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-
-void TemplateTable::if_acmp(Condition cc) {
-  transition(atos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ pop_ptr(rdx);
-  __ cmpptr(rdx, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-
-void TemplateTable::ret() {
-  transition(vtos, vtos);
-  locals_index(rbx);
-  __ movptr(rbx, iaddress(rbx));                   // get return bci, compute return bcp
-  __ profile_ret(rbx, rcx);
-  __ get_method(rax);
-  __ movptr(rsi, Address(rax, Method::const_offset()));
-  __ lea(rsi, Address(rsi, rbx, Address::times_1,
-                      ConstMethod::codes_offset()));
-  __ dispatch_next(vtos);
-}
-
-
-void TemplateTable::wide_ret() {
-  transition(vtos, vtos);
-  locals_index_wide(rbx);
-  __ movptr(rbx, iaddress(rbx));                   // get return bci, compute return bcp
-  __ profile_ret(rbx, rcx);
-  __ get_method(rax);
-  __ movptr(rsi, Address(rax, Method::const_offset()));
-  __ lea(rsi, Address(rsi, rbx, Address::times_1, ConstMethod::codes_offset()));
-  __ dispatch_next(vtos);
-}
-
-
-void TemplateTable::tableswitch() {
-  Label default_case, continue_execution;
-  transition(itos, vtos);
-  // align rsi
-  __ lea(rbx, at_bcp(wordSize));
-  __ andptr(rbx, -wordSize);
-  // load lo & hi
-  __ movl(rcx, Address(rbx, 1 * wordSize));
-  __ movl(rdx, Address(rbx, 2 * wordSize));
-  __ bswapl(rcx);
-  __ bswapl(rdx);
-  // check against lo & hi
-  __ cmpl(rax, rcx);
-  __ jccb(Assembler::less, default_case);
-  __ cmpl(rax, rdx);
-  __ jccb(Assembler::greater, default_case);
-  // lookup dispatch offset
-  __ subl(rax, rcx);
-  __ movl(rdx, Address(rbx, rax, Address::times_4, 3 * BytesPerInt));
-  __ profile_switch_case(rax, rbx, rcx);
-  // continue execution
-  __ bind(continue_execution);
-  __ bswapl(rdx);
-  __ load_unsigned_byte(rbx, Address(rsi, rdx, Address::times_1));
-  __ addptr(rsi, rdx);
-  __ dispatch_only(vtos);
-  // handle default
-  __ bind(default_case);
-  __ profile_switch_default(rax);
-  __ movl(rdx, Address(rbx, 0));
-  __ jmp(continue_execution);
-}
-
-
-void TemplateTable::lookupswitch() {
-  transition(itos, itos);
-  __ stop("lookupswitch bytecode should have been rewritten");
-}
-
-
-void TemplateTable::fast_linearswitch() {
-  transition(itos, vtos);
-  Label loop_entry, loop, found, continue_execution;
-  // bswapl rax, so we can avoid bswapping the table entries
-  __ bswapl(rax);
-  // align rsi
-  __ lea(rbx, at_bcp(wordSize));                // btw: should be able to get rid of this instruction (change offsets below)
-  __ andptr(rbx, -wordSize);
-  // set counter
-  __ movl(rcx, Address(rbx, wordSize));
-  __ bswapl(rcx);
-  __ jmpb(loop_entry);
-  // table search
-  __ bind(loop);
-  __ cmpl(rax, Address(rbx, rcx, Address::times_8, 2 * wordSize));
-  __ jccb(Assembler::equal, found);
-  __ bind(loop_entry);
-  __ decrementl(rcx);
-  __ jcc(Assembler::greaterEqual, loop);
-  // default case
-  __ profile_switch_default(rax);
-  __ movl(rdx, Address(rbx, 0));
-  __ jmpb(continue_execution);
-  // entry found -> get offset
-  __ bind(found);
-  __ movl(rdx, Address(rbx, rcx, Address::times_8, 3 * wordSize));
-  __ profile_switch_case(rcx, rax, rbx);
-  // continue execution
-  __ bind(continue_execution);
-  __ bswapl(rdx);
-  __ load_unsigned_byte(rbx, Address(rsi, rdx, Address::times_1));
-  __ addptr(rsi, rdx);
-  __ dispatch_only(vtos);
-}
-
-
-void TemplateTable::fast_binaryswitch() {
-  transition(itos, vtos);
-  // Implementation using the following core algorithm:
-  //
-  // int binary_search(int key, LookupswitchPair* array, int n) {
-  //   // Binary search according to "Methodik des Programmierens" by
-  //   // Edsger W. Dijkstra and W.H.J. Feijen, Addison Wesley Germany 1985.
-  //   int i = 0;
-  //   int j = n;
-  //   while (i+1 < j) {
-  //     // invariant P: 0 <= i < j <= n and (a[i] <= key < a[j] or Q)
-  //     // with      Q: for all i: 0 <= i < n: key < a[i]
-  //     // where a stands for the array and assuming that the (inexisting)
-  //     // element a[n] is infinitely big.
-  //     int h = (i + j) >> 1;
-  //     // i < h < j
-  //     if (key < array[h].fast_match()) {
-  //       j = h;
-  //     } else {
-  //       i = h;
-  //     }
-  //   }
-  //   // R: a[i] <= key < a[i+1] or Q
-  //   // (i.e., if key is within array, i is the correct index)
-  //   return i;
-  // }
-
-  // register allocation
-  const Register key   = rax;                    // already set (tosca)
-  const Register array = rbx;
-  const Register i     = rcx;
-  const Register j     = rdx;
-  const Register h     = rdi;                    // needs to be restored
-  const Register temp  = rsi;
-  // setup array
-  __ save_bcp();
-
-  __ lea(array, at_bcp(3*wordSize));             // btw: should be able to get rid of this instruction (change offsets below)
-  __ andptr(array, -wordSize);
-  // initialize i & j
-  __ xorl(i, i);                                 // i = 0;
-  __ movl(j, Address(array, -wordSize));         // j = length(array);
-  // Convert j into native byteordering
-  __ bswapl(j);
-  // and start
-  Label entry;
-  __ jmp(entry);
-
-  // binary search loop
-  { Label loop;
-    __ bind(loop);
-    // int h = (i + j) >> 1;
-    __ leal(h, Address(i, j, Address::times_1)); // h = i + j;
-    __ sarl(h, 1);                               // h = (i + j) >> 1;
-    // if (key < array[h].fast_match()) {
-    //   j = h;
-    // } else {
-    //   i = h;
-    // }
-    // Convert array[h].match to native byte-ordering before compare
-    __ movl(temp, Address(array, h, Address::times_8, 0*wordSize));
-    __ bswapl(temp);
-    __ cmpl(key, temp);
-    // j = h if (key <  array[h].fast_match())
-    __ cmov32(Assembler::less        , j, h);
-    // i = h if (key >= array[h].fast_match())
-    __ cmov32(Assembler::greaterEqual, i, h);
-    // while (i+1 < j)
-    __ bind(entry);
-    __ leal(h, Address(i, 1));                   // i+1
-    __ cmpl(h, j);                               // i+1 < j
-    __ jcc(Assembler::less, loop);
-  }
-
-  // end of binary search, result index is i (must check again!)
-  Label default_case;
-  // Convert array[i].match to native byte-ordering before compare
-  __ movl(temp, Address(array, i, Address::times_8, 0*wordSize));
-  __ bswapl(temp);
-  __ cmpl(key, temp);
-  __ jcc(Assembler::notEqual, default_case);
-
-  // entry found -> j = offset
-  __ movl(j , Address(array, i, Address::times_8, 1*wordSize));
-  __ profile_switch_case(i, key, array);
-  __ bswapl(j);
-  LP64_ONLY(__ movslq(j, j));
-  __ restore_bcp();
-  __ restore_locals();                           // restore rdi
-  __ load_unsigned_byte(rbx, Address(rsi, j, Address::times_1));
-
-  __ addptr(rsi, j);
-  __ dispatch_only(vtos);
-
-  // default case -> j = default offset
-  __ bind(default_case);
-  __ profile_switch_default(i);
-  __ movl(j, Address(array, -2*wordSize));
-  __ bswapl(j);
-  LP64_ONLY(__ movslq(j, j));
-  __ restore_bcp();
-  __ restore_locals();                           // restore rdi
-  __ load_unsigned_byte(rbx, Address(rsi, j, Address::times_1));
-  __ addptr(rsi, j);
-  __ dispatch_only(vtos);
-}
-
-
-void TemplateTable::_return(TosState state) {
-  transition(state, state);
-  assert(_desc->calls_vm(), "inconsistent calls_vm information"); // call in remove_activation
-
-  if (_desc->bytecode() == Bytecodes::_return_register_finalizer) {
-    assert(state == vtos, "only valid state");
-    __ movptr(rax, aaddress(0));
-    __ load_klass(rdi, rax);
-    __ movl(rdi, Address(rdi, Klass::access_flags_offset()));
-    __ testl(rdi, JVM_ACC_HAS_FINALIZER);
-    Label skip_register_finalizer;
-    __ jcc(Assembler::zero, skip_register_finalizer);
-
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::register_finalizer), rax);
-
-    __ bind(skip_register_finalizer);
-  }
-
-  __ remove_activation(state, rsi);
-  __ jmp(rsi);
-}
-
-
-// ----------------------------------------------------------------------------
-// Volatile variables demand their effects be made known to all CPU's in
-// order.  Store buffers on most chips allow reads & writes to reorder; the
-// JMM's ReadAfterWrite.java test fails in -Xint mode without some kind of
-// memory barrier (i.e., it's not sufficient that the interpreter does not
-// reorder volatile references, the hardware also must not reorder them).
-//
-// According to the new Java Memory Model (JMM):
-// (1) All volatiles are serialized wrt to each other.
-// ALSO reads & writes act as aquire & release, so:
-// (2) A read cannot let unrelated NON-volatile memory refs that happen after
-// the read float up to before the read.  It's OK for non-volatile memory refs
-// that happen before the volatile read to float down below it.
-// (3) Similar a volatile write cannot let unrelated NON-volatile memory refs
-// that happen BEFORE the write float down to after the write.  It's OK for
-// non-volatile memory refs that happen after the volatile write to float up
-// before it.
-//
-// We only put in barriers around volatile refs (they are expensive), not
-// _between_ memory refs (that would require us to track the flavor of the
-// previous memory refs).  Requirements (2) and (3) require some barriers
-// before volatile stores and after volatile loads.  These nearly cover
-// requirement (1) but miss the volatile-store-volatile-load case.  This final
-// case is placed after volatile-stores although it could just as well go
-// before volatile-loads.
-void TemplateTable::volatile_barrier(Assembler::Membar_mask_bits order_constraint ) {
-  // Helper function to insert a is-volatile test and memory barrier
-  if( !os::is_MP() ) return;    // Not needed on single CPU
-  __ membar(order_constraint);
-}
-
-void TemplateTable::resolve_cache_and_index(int byte_no,
-                                            Register Rcache,
-                                            Register index,
-                                            size_t index_size) {
-  const Register temp = rbx;
-  assert_different_registers(Rcache, index, temp);
-
-  Label resolved;
-    assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
-    __ get_cache_and_index_and_bytecode_at_bcp(Rcache, index, temp, byte_no, 1, index_size);
-    __ cmpl(temp, (int) bytecode());  // have we resolved this bytecode?
-    __ jcc(Assembler::equal, resolved);
-
-  // resolve first time through
-  address entry;
-  switch (bytecode()) {
-    case Bytecodes::_getstatic      : // fall through
-    case Bytecodes::_putstatic      : // fall through
-    case Bytecodes::_getfield       : // fall through
-    case Bytecodes::_putfield       : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_get_put);        break;
-    case Bytecodes::_invokevirtual  : // fall through
-    case Bytecodes::_invokespecial  : // fall through
-    case Bytecodes::_invokestatic   : // fall through
-    case Bytecodes::_invokeinterface: entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invoke);         break;
-    case Bytecodes::_invokehandle   : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokehandle);   break;
-    case Bytecodes::_invokedynamic  : entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokedynamic);  break;
-    default:
-      fatal(err_msg("unexpected bytecode: %s", Bytecodes::name(bytecode())));
-      break;
-  }
-  __ movl(temp, (int)bytecode());
-  __ call_VM(noreg, entry, temp);
-  // Update registers with resolved info
-  __ get_cache_and_index_at_bcp(Rcache, index, 1, index_size);
-  __ bind(resolved);
-}
-
-
-// The cache and index registers must be set before call
-void TemplateTable::load_field_cp_cache_entry(Register obj,
-                                              Register cache,
-                                              Register index,
-                                              Register off,
-                                              Register flags,
-                                              bool is_static = false) {
-  assert_different_registers(cache, index, flags, off);
-
-  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
-  // Field offset
-  __ movptr(off, Address(cache, index, Address::times_ptr,
-                         in_bytes(cp_base_offset + ConstantPoolCacheEntry::f2_offset())));
-  // Flags
-  __ movl(flags, Address(cache, index, Address::times_ptr,
-           in_bytes(cp_base_offset + ConstantPoolCacheEntry::flags_offset())));
-
-  // klass overwrite register
-  if (is_static) {
-    __ movptr(obj, Address(cache, index, Address::times_ptr,
-                           in_bytes(cp_base_offset + ConstantPoolCacheEntry::f1_offset())));
-    const int mirror_offset = in_bytes(Klass::java_mirror_offset());
-    __ movptr(obj, Address(obj, mirror_offset));
-  }
-}
-
-void TemplateTable::load_invoke_cp_cache_entry(int byte_no,
-                                               Register method,
-                                               Register itable_index,
-                                               Register flags,
-                                               bool is_invokevirtual,
-                                               bool is_invokevfinal, /*unused*/
-                                               bool is_invokedynamic) {
-  // setup registers
-  const Register cache = rcx;
-  const Register index = rdx;
-  assert_different_registers(method, flags);
-  assert_different_registers(method, cache, index);
-  assert_different_registers(itable_index, flags);
-  assert_different_registers(itable_index, cache, index);
-  // determine constant pool cache field offsets
-  assert(is_invokevirtual == (byte_no == f2_byte), "is_invokevirtual flag redundant");
-  const int method_offset = in_bytes(
-    ConstantPoolCache::base_offset() +
-      ((byte_no == f2_byte)
-       ? ConstantPoolCacheEntry::f2_offset()
-       : ConstantPoolCacheEntry::f1_offset()));
-  const int flags_offset = in_bytes(ConstantPoolCache::base_offset() +
-                                    ConstantPoolCacheEntry::flags_offset());
-  // access constant pool cache fields
-  const int index_offset = in_bytes(ConstantPoolCache::base_offset() +
-                                    ConstantPoolCacheEntry::f2_offset());
-
-  size_t index_size = (is_invokedynamic ? sizeof(u4) : sizeof(u2));
-  resolve_cache_and_index(byte_no, cache, index, index_size);
-    __ movptr(method, Address(cache, index, Address::times_ptr, method_offset));
-
-  if (itable_index != noreg) {
-    __ movptr(itable_index, Address(cache, index, Address::times_ptr, index_offset));
-  }
-  __ movl(flags, Address(cache, index, Address::times_ptr, flags_offset));
-}
-
-
-// The registers cache and index expected to be set before call.
-// Correct values of the cache and index registers are preserved.
-void TemplateTable::jvmti_post_field_access(Register cache,
-                                            Register index,
-                                            bool is_static,
-                                            bool has_tos) {
-  if (JvmtiExport::can_post_field_access()) {
-    // Check to see if a field access watch has been set before we take
-    // the time to call into the VM.
-    Label L1;
-    assert_different_registers(cache, index, rax);
-    __ mov32(rax, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
-    __ testl(rax,rax);
-    __ jcc(Assembler::zero, L1);
-
-    // cache entry pointer
-    __ addptr(cache, in_bytes(ConstantPoolCache::base_offset()));
-    __ shll(index, LogBytesPerWord);
-    __ addptr(cache, index);
-    if (is_static) {
-      __ xorptr(rax, rax);      // NULL object reference
-    } else {
-      __ pop(atos);         // Get the object
-      __ verify_oop(rax);
-      __ push(atos);        // Restore stack state
-    }
-    // rax,:   object pointer or NULL
-    // cache: cache entry pointer
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_access),
-               rax, cache);
-    __ get_cache_and_index_at_bcp(cache, index, 1);
-    __ bind(L1);
-  }
-}
-
-void TemplateTable::pop_and_check_object(Register r) {
-  __ pop_ptr(r);
-  __ null_check(r);  // for field access must check obj.
-  __ verify_oop(r);
-}
-
-void TemplateTable::getfield_or_static(int byte_no, bool is_static) {
-  transition(vtos, vtos);
-
-  const Register cache = rcx;
-  const Register index = rdx;
-  const Register obj   = rcx;
-  const Register off   = rbx;
-  const Register flags = rax;
-
-  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
-  jvmti_post_field_access(cache, index, is_static, false);
-  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
-
-  if (!is_static) pop_and_check_object(obj);
-
-  const Address lo(obj, off, Address::times_1, 0*wordSize);
-  const Address hi(obj, off, Address::times_1, 1*wordSize);
-
-  Label Done, notByte, notInt, notShort, notChar, notLong, notFloat, notObj, notDouble;
-
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-  assert(btos == 0, "change code, btos != 0");
-  // btos
-  __ andptr(flags, ConstantPoolCacheEntry::tos_state_mask);
-  __ jcc(Assembler::notZero, notByte);
-
-  __ load_signed_byte(rax, lo );
-  __ push(btos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_bgetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notByte);
-  // itos
-  __ cmpl(flags, itos );
-  __ jcc(Assembler::notEqual, notInt);
-
-  __ movl(rax, lo );
-  __ push(itos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_igetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notInt);
-  // atos
-  __ cmpl(flags, atos );
-  __ jcc(Assembler::notEqual, notObj);
-
-  __ movl(rax, lo );
-  __ push(atos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_agetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notObj);
-  // ctos
-  __ cmpl(flags, ctos );
-  __ jcc(Assembler::notEqual, notChar);
-
-  __ load_unsigned_short(rax, lo );
-  __ push(ctos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_cgetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notChar);
-  // stos
-  __ cmpl(flags, stos );
-  __ jcc(Assembler::notEqual, notShort);
-
-  __ load_signed_short(rax, lo );
-  __ push(stos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_sgetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notShort);
-  // ltos
-  __ cmpl(flags, ltos );
-  __ jcc(Assembler::notEqual, notLong);
-
-  // Generate code as if volatile.  There just aren't enough registers to
-  // save that information and this code is faster than the test.
-  __ fild_d(lo);                // Must load atomically
-  __ subptr(rsp,2*wordSize);    // Make space for store
-  __ fistp_d(Address(rsp,0));
-  __ pop(rax);
-  __ pop(rdx);
-
-  __ push(ltos);
-  // Don't rewrite to _fast_lgetfield for potential volatile case.
-  __ jmp(Done);
-
-  __ bind(notLong);
-  // ftos
-  __ cmpl(flags, ftos );
-  __ jcc(Assembler::notEqual, notFloat);
-
-  __ fld_s(lo);
-  __ push(ftos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_fgetfield, rcx, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notFloat);
-  // dtos
-  __ cmpl(flags, dtos );
-  __ jcc(Assembler::notEqual, notDouble);
-
-  __ fld_d(lo);
-  __ push(dtos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_dgetfield, rcx, rbx);
-  }
-  __ jmpb(Done);
-
-  __ bind(notDouble);
-
-  __ stop("Bad state");
-
-  __ bind(Done);
-  // Doug Lea believes this is not needed with current Sparcs (TSO) and Intel (PSO).
-  // volatile_barrier( );
-}
-
-
-void TemplateTable::getfield(int byte_no) {
-  getfield_or_static(byte_no, false);
-}
-
-
-void TemplateTable::getstatic(int byte_no) {
-  getfield_or_static(byte_no, true);
-}
-
-// The registers cache and index expected to be set before call.
-// The function may destroy various registers, just not the cache and index registers.
-void TemplateTable::jvmti_post_field_mod(Register cache, Register index, bool is_static) {
-
-  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
-
-  if (JvmtiExport::can_post_field_modification()) {
-    // Check to see if a field modification watch has been set before we take
-    // the time to call into the VM.
-    Label L1;
-    assert_different_registers(cache, index, rax);
-    __ mov32(rax, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
-    __ testl(rax, rax);
-    __ jcc(Assembler::zero, L1);
-
-    // The cache and index registers have been already set.
-    // This allows to eliminate this call but the cache and index
-    // registers have to be correspondingly used after this line.
-    __ get_cache_and_index_at_bcp(rax, rdx, 1);
-
-    if (is_static) {
-      // Life is simple.  Null out the object pointer.
-      __ xorptr(rbx, rbx);
-    } else {
-      // Life is harder. The stack holds the value on top, followed by the object.
-      // We don't know the size of the value, though; it could be one or two words
-      // depending on its type. As a result, we must find the type to determine where
-      // the object is.
-      Label two_word, valsize_known;
-      __ movl(rcx, Address(rax, rdx, Address::times_ptr, in_bytes(cp_base_offset +
-                                   ConstantPoolCacheEntry::flags_offset())));
-      __ mov(rbx, rsp);
-      __ shrl(rcx, ConstantPoolCacheEntry::tos_state_shift);
-      // Make sure we don't need to mask rcx after the above shift
-      ConstantPoolCacheEntry::verify_tos_state_shift();
-      __ cmpl(rcx, ltos);
-      __ jccb(Assembler::equal, two_word);
-      __ cmpl(rcx, dtos);
-      __ jccb(Assembler::equal, two_word);
-      __ addptr(rbx, Interpreter::expr_offset_in_bytes(1)); // one word jvalue (not ltos, dtos)
-      __ jmpb(valsize_known);
-
-      __ bind(two_word);
-      __ addptr(rbx, Interpreter::expr_offset_in_bytes(2)); // two words jvalue
-
-      __ bind(valsize_known);
-      // setup object pointer
-      __ movptr(rbx, Address(rbx, 0));
-    }
-    // cache entry pointer
-    __ addptr(rax, in_bytes(cp_base_offset));
-    __ shll(rdx, LogBytesPerWord);
-    __ addptr(rax, rdx);
-    // object (tos)
-    __ mov(rcx, rsp);
-    // rbx,: object pointer set up above (NULL if static)
-    // rax,: cache entry pointer
-    // rcx: jvalue object on the stack
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_modification),
-               rbx, rax, rcx);
-    __ get_cache_and_index_at_bcp(cache, index, 1);
-    __ bind(L1);
-  }
-}
-
-
-void TemplateTable::putfield_or_static(int byte_no, bool is_static) {
-  transition(vtos, vtos);
-
-  const Register cache = rcx;
-  const Register index = rdx;
-  const Register obj   = rcx;
-  const Register off   = rbx;
-  const Register flags = rax;
-
-  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
-  jvmti_post_field_mod(cache, index, is_static);
-  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
-
-  // Doug Lea believes this is not needed with current Sparcs (TSO) and Intel (PSO).
-  // volatile_barrier( );
-
-  Label notVolatile, Done;
-  __ movl(rdx, flags);
-  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  __ andl(rdx, 0x1);
-
-  // field addresses
-  const Address lo(obj, off, Address::times_1, 0*wordSize);
-  const Address hi(obj, off, Address::times_1, 1*wordSize);
-
-  Label notByte, notInt, notShort, notChar, notLong, notFloat, notObj, notDouble;
-
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-  assert(btos == 0, "change code, btos != 0");
-  __ andl(flags, ConstantPoolCacheEntry::tos_state_mask);
-  __ jcc(Assembler::notZero, notByte);
-
-  // btos
-  {
-    __ pop(btos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movb(lo, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_bputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notByte);
-  __ cmpl(flags, itos);
-  __ jcc(Assembler::notEqual, notInt);
-
-  // itos
-  {
-    __ pop(itos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movl(lo, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_iputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notInt);
-  __ cmpl(flags, atos);
-  __ jcc(Assembler::notEqual, notObj);
-
-  // atos
-  {
-    __ pop(atos);
-    if (!is_static) pop_and_check_object(obj);
-    do_oop_store(_masm, lo, rax, _bs->kind(), false);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_aputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notObj);
-  __ cmpl(flags, ctos);
-  __ jcc(Assembler::notEqual, notChar);
-
-  // ctos
-  {
-    __ pop(ctos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movw(lo, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_cputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notChar);
-  __ cmpl(flags, stos);
-  __ jcc(Assembler::notEqual, notShort);
-
-  // stos
-  {
-    __ pop(stos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movw(lo, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_sputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notShort);
-  __ cmpl(flags, ltos);
-  __ jcc(Assembler::notEqual, notLong);
-
-  // ltos
-  {
-    Label notVolatileLong;
-    __ testl(rdx, rdx);
-    __ jcc(Assembler::zero, notVolatileLong);
-
-    __ pop(ltos);  // overwrites rdx, do this after testing volatile.
-    if (!is_static) pop_and_check_object(obj);
-
-    // Replace with real volatile test
-    __ push(rdx);
-    __ push(rax);                 // Must update atomically with FIST
-    __ fild_d(Address(rsp,0));    // So load into FPU register
-    __ fistp_d(lo);               // and put into memory atomically
-    __ addptr(rsp, 2*wordSize);
-    // volatile_barrier();
-    volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
-                                                 Assembler::StoreStore));
-    // Don't rewrite volatile version
-    __ jmp(notVolatile);
-
-    __ bind(notVolatileLong);
-
-    __ pop(ltos);  // overwrites rdx
-    if (!is_static) pop_and_check_object(obj);
-    NOT_LP64(__ movptr(hi, rdx));
-    __ movptr(lo, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_lputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(notVolatile);
-  }
-
-  __ bind(notLong);
-  __ cmpl(flags, ftos);
-  __ jcc(Assembler::notEqual, notFloat);
-
-  // ftos
-  {
-    __ pop(ftos);
-    if (!is_static) pop_and_check_object(obj);
-    __ fstp_s(lo);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_fputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notFloat);
-#ifdef ASSERT
-  __ cmpl(flags, dtos);
-  __ jcc(Assembler::notEqual, notDouble);
-#endif
-
-  // dtos
-  {
-    __ pop(dtos);
-    if (!is_static) pop_and_check_object(obj);
-    __ fstp_d(lo);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_dputfield, rcx, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-#ifdef ASSERT
-  __ bind(notDouble);
-  __ stop("Bad state");
-#endif
-
-  __ bind(Done);
-
-  // Check for volatile store
-  __ testl(rdx, rdx);
-  __ jcc(Assembler::zero, notVolatile);
-  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
-                                               Assembler::StoreStore));
-  __ bind(notVolatile);
-}
-
-
-void TemplateTable::putfield(int byte_no) {
-  putfield_or_static(byte_no, false);
-}
-
-
-void TemplateTable::putstatic(int byte_no) {
-  putfield_or_static(byte_no, true);
-}
-
-void TemplateTable::jvmti_post_fast_field_mod() {
-  if (JvmtiExport::can_post_field_modification()) {
-    // Check to see if a field modification watch has been set before we take
-    // the time to call into the VM.
-    Label L2;
-     __ mov32(rcx, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
-     __ testl(rcx,rcx);
-     __ jcc(Assembler::zero, L2);
-     __ pop_ptr(rbx);               // copy the object pointer from tos
-     __ verify_oop(rbx);
-     __ push_ptr(rbx);              // put the object pointer back on tos
-
-     // Save tos values before call_VM() clobbers them. Since we have
-     // to do it for every data type, we use the saved values as the
-     // jvalue object.
-     switch (bytecode()) {          // load values into the jvalue object
-     case Bytecodes::_fast_aputfield: __ push_ptr(rax); break;
-     case Bytecodes::_fast_bputfield: // fall through
-     case Bytecodes::_fast_sputfield: // fall through
-     case Bytecodes::_fast_cputfield: // fall through
-     case Bytecodes::_fast_iputfield: __ push_i(rax); break;
-     case Bytecodes::_fast_dputfield: __ push_d(); break;
-     case Bytecodes::_fast_fputfield: __ push_f(); break;
-     case Bytecodes::_fast_lputfield: __ push_l(rax); break;
-
-     default:
-       ShouldNotReachHere();
-     }
-     __ mov(rcx, rsp);              // points to jvalue on the stack
-     // access constant pool cache entry
-     __ get_cache_entry_pointer_at_bcp(rax, rdx, 1);
-     __ verify_oop(rbx);
-     // rbx,: object pointer copied above
-     // rax,: cache entry pointer
-     // rcx: jvalue object on the stack
-     __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_modification), rbx, rax, rcx);
-
-     switch (bytecode()) {             // restore tos values
-     case Bytecodes::_fast_aputfield: __ pop_ptr(rax); break;
-     case Bytecodes::_fast_bputfield: // fall through
-     case Bytecodes::_fast_sputfield: // fall through
-     case Bytecodes::_fast_cputfield: // fall through
-     case Bytecodes::_fast_iputfield: __ pop_i(rax); break;
-     case Bytecodes::_fast_dputfield: __ pop_d(); break;
-     case Bytecodes::_fast_fputfield: __ pop_f(); break;
-     case Bytecodes::_fast_lputfield: __ pop_l(rax); break;
-     }
-     __ bind(L2);
-  }
-}
-
-void TemplateTable::fast_storefield(TosState state) {
-  transition(state, vtos);
-
-  ByteSize base = ConstantPoolCache::base_offset();
-
-  jvmti_post_fast_field_mod();
-
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
-
-  // test for volatile with rdx but rdx is tos register for lputfield.
-  if (bytecode() == Bytecodes::_fast_lputfield) __ push(rdx);
-  __ movl(rdx, Address(rcx, rbx, Address::times_ptr, in_bytes(base +
-                       ConstantPoolCacheEntry::flags_offset())));
-
-  // replace index with field offset from cache entry
-  __ movptr(rbx, Address(rcx, rbx, Address::times_ptr, in_bytes(base + ConstantPoolCacheEntry::f2_offset())));
-
-  // Doug Lea believes this is not needed with current Sparcs (TSO) and Intel (PSO).
-  // volatile_barrier( );
-
-  Label notVolatile, Done;
-  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  __ andl(rdx, 0x1);
-  // Check for volatile store
-  __ testl(rdx, rdx);
-  __ jcc(Assembler::zero, notVolatile);
-
-  if (bytecode() == Bytecodes::_fast_lputfield) __ pop(rdx);
-
-  // Get object from stack
-  pop_and_check_object(rcx);
-
-  // field addresses
-  const Address lo(rcx, rbx, Address::times_1, 0*wordSize);
-  const Address hi(rcx, rbx, Address::times_1, 1*wordSize);
-
-  // access field
-  switch (bytecode()) {
-    case Bytecodes::_fast_bputfield: __ movb(lo, rax); break;
-    case Bytecodes::_fast_sputfield: // fall through
-    case Bytecodes::_fast_cputfield: __ movw(lo, rax); break;
-    case Bytecodes::_fast_iputfield: __ movl(lo, rax); break;
-    case Bytecodes::_fast_lputfield:
-      NOT_LP64(__ movptr(hi, rdx));
-      __ movptr(lo, rax);
-      break;
-    case Bytecodes::_fast_fputfield: __ fstp_s(lo); break;
-    case Bytecodes::_fast_dputfield: __ fstp_d(lo); break;
-    case Bytecodes::_fast_aputfield: {
-      do_oop_store(_masm, lo, rax, _bs->kind(), false);
-      break;
-    }
-    default:
-      ShouldNotReachHere();
-  }
-
-  Label done;
-  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
-                                               Assembler::StoreStore));
-  // Barriers are so large that short branch doesn't reach!
-  __ jmp(done);
-
-  // Same code as above, but don't need rdx to test for volatile.
-  __ bind(notVolatile);
-
-  if (bytecode() == Bytecodes::_fast_lputfield) __ pop(rdx);
-
-  // Get object from stack
-  pop_and_check_object(rcx);
-
-  // access field
-  switch (bytecode()) {
-    case Bytecodes::_fast_bputfield: __ movb(lo, rax); break;
-    case Bytecodes::_fast_sputfield: // fall through
-    case Bytecodes::_fast_cputfield: __ movw(lo, rax); break;
-    case Bytecodes::_fast_iputfield: __ movl(lo, rax); break;
-    case Bytecodes::_fast_lputfield:
-      NOT_LP64(__ movptr(hi, rdx));
-      __ movptr(lo, rax);
-      break;
-    case Bytecodes::_fast_fputfield: __ fstp_s(lo); break;
-    case Bytecodes::_fast_dputfield: __ fstp_d(lo); break;
-    case Bytecodes::_fast_aputfield: {
-      do_oop_store(_masm, lo, rax, _bs->kind(), false);
-      break;
-    }
-    default:
-      ShouldNotReachHere();
-  }
-  __ bind(done);
-}
-
-
-void TemplateTable::fast_accessfield(TosState state) {
-  transition(atos, state);
-
-  // do the JVMTI work here to avoid disturbing the register state below
-  if (JvmtiExport::can_post_field_access()) {
-    // Check to see if a field access watch has been set before we take
-    // the time to call into the VM.
-    Label L1;
-    __ mov32(rcx, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
-    __ testl(rcx,rcx);
-    __ jcc(Assembler::zero, L1);
-    // access constant pool cache entry
-    __ get_cache_entry_pointer_at_bcp(rcx, rdx, 1);
-    __ push_ptr(rax);  // save object pointer before call_VM() clobbers it
-    __ verify_oop(rax);
-    // rax,: object pointer copied above
-    // rcx: cache entry pointer
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::post_field_access), rax, rcx);
-    __ pop_ptr(rax);   // restore object pointer
-    __ bind(L1);
-  }
-
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
-  // replace index with field offset from cache entry
-  __ movptr(rbx, Address(rcx,
-                         rbx,
-                         Address::times_ptr,
-                         in_bytes(ConstantPoolCache::base_offset() + ConstantPoolCacheEntry::f2_offset())));
-
-
-  // rax,: object
-  __ verify_oop(rax);
-  __ null_check(rax);
-  // field addresses
-  const Address lo = Address(rax, rbx, Address::times_1, 0*wordSize);
-  const Address hi = Address(rax, rbx, Address::times_1, 1*wordSize);
-
-  // access field
-  switch (bytecode()) {
-    case Bytecodes::_fast_bgetfield: __ movsbl(rax, lo );                 break;
-    case Bytecodes::_fast_sgetfield: __ load_signed_short(rax, lo );      break;
-    case Bytecodes::_fast_cgetfield: __ load_unsigned_short(rax, lo );    break;
-    case Bytecodes::_fast_igetfield: __ movl(rax, lo);                    break;
-    case Bytecodes::_fast_lgetfield: __ stop("should not be rewritten");  break;
-    case Bytecodes::_fast_fgetfield: __ fld_s(lo);                        break;
-    case Bytecodes::_fast_dgetfield: __ fld_d(lo);                        break;
-    case Bytecodes::_fast_agetfield: __ movptr(rax, lo); __ verify_oop(rax); break;
-    default:
-      ShouldNotReachHere();
-  }
-
-  // Doug Lea believes this is not needed with current Sparcs(TSO) and Intel(PSO)
-  // volatile_barrier( );
-}
-
-void TemplateTable::fast_xaccess(TosState state) {
-  transition(vtos, state);
-  // get receiver
-  __ movptr(rax, aaddress(0));
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rdx, 2);
-  __ movptr(rbx, Address(rcx,
-                         rdx,
-                         Address::times_ptr,
-                         in_bytes(ConstantPoolCache::base_offset() + ConstantPoolCacheEntry::f2_offset())));
-  // make sure exception is reported in correct bcp range (getfield is next instruction)
-  __ increment(rsi);
-  __ null_check(rax);
-  const Address lo = Address(rax, rbx, Address::times_1, 0*wordSize);
-  if (state == itos) {
-    __ movl(rax, lo);
-  } else if (state == atos) {
-    __ movptr(rax, lo);
-    __ verify_oop(rax);
-  } else if (state == ftos) {
-    __ fld_s(lo);
-  } else {
-    ShouldNotReachHere();
-  }
-  __ decrement(rsi);
-}
-
-
-
-//----------------------------------------------------------------------------------------------------
-// Calls
-
-void TemplateTable::count_calls(Register method, Register temp) {
-  // implemented elsewhere
-  ShouldNotReachHere();
-}
-
-
-void TemplateTable::prepare_invoke(int byte_no,
-                                   Register method,  // linked method (or i-klass)
-                                   Register index,   // itable index, MethodType, etc.
-                                   Register recv,    // if caller wants to see it
-                                   Register flags    // if caller wants to test it
-                                   ) {
-  // determine flags
-  const Bytecodes::Code code = bytecode();
-  const bool is_invokeinterface  = code == Bytecodes::_invokeinterface;
-  const bool is_invokedynamic    = code == Bytecodes::_invokedynamic;
-  const bool is_invokehandle     = code == Bytecodes::_invokehandle;
-  const bool is_invokevirtual    = code == Bytecodes::_invokevirtual;
-  const bool is_invokespecial    = code == Bytecodes::_invokespecial;
-  const bool load_receiver       = (recv  != noreg);
-  const bool save_flags          = (flags != noreg);
-  assert(load_receiver == (code != Bytecodes::_invokestatic && code != Bytecodes::_invokedynamic), "");
-  assert(save_flags    == (is_invokeinterface || is_invokevirtual), "need flags for vfinal");
-  assert(flags == noreg || flags == rdx, "");
-  assert(recv  == noreg || recv  == rcx, "");
-
-  // setup registers & access constant pool cache
-  if (recv  == noreg)  recv  = rcx;
-  if (flags == noreg)  flags = rdx;
-  assert_different_registers(method, index, recv, flags);
-
-  // save 'interpreter return address'
-  __ save_bcp();
-
-  load_invoke_cp_cache_entry(byte_no, method, index, flags, is_invokevirtual, false, is_invokedynamic);
-
-  // maybe push appendix to arguments (just before return address)
-  if (is_invokedynamic || is_invokehandle) {
-    Label L_no_push;
-    __ testl(flags, (1 << ConstantPoolCacheEntry::has_appendix_shift));
-    __ jccb(Assembler::zero, L_no_push);
-    // Push the appendix as a trailing parameter.
-    // This must be done before we get the receiver,
-    // since the parameter_size includes it.
-    __ push(rbx);
-    __ mov(rbx, index);
-    assert(ConstantPoolCacheEntry::_indy_resolved_references_appendix_offset == 0, "appendix expected at index+0");
-    __ load_resolved_reference_at_index(index, rbx);
-    __ pop(rbx);
-    __ push(index);  // push appendix (MethodType, CallSite, etc.)
-    __ bind(L_no_push);
-  }
-
-  // load receiver if needed (note: no return address pushed yet)
-  if (load_receiver) {
-    __ movl(recv, flags);
-    __ andl(recv, ConstantPoolCacheEntry::parameter_size_mask);
-    const int no_return_pc_pushed_yet = -1;  // argument slot correction before we push return address
-    const int receiver_is_at_end      = -1;  // back off one slot to get receiver
-    Address recv_addr = __ argument_address(recv, no_return_pc_pushed_yet + receiver_is_at_end);
-    __ movptr(recv, recv_addr);
-    __ verify_oop(recv);
-  }
-
-  if (save_flags) {
-    __ mov(rsi, flags);
-  }
-
-  // compute return type
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-  // Make sure we don't need to mask flags after the above shift
-  ConstantPoolCacheEntry::verify_tos_state_shift();
-  // load return address
-  {
-    const address table_addr = (address) Interpreter::invoke_return_entry_table_for(code);
-    ExternalAddress table(table_addr);
-    __ movptr(flags, ArrayAddress(table, Address(noreg, flags, Address::times_ptr)));
-  }
-
-  // push return address
-  __ push(flags);
-
-  // Restore flags value from the constant pool cache, and restore rsi
-  // for later null checks.  rsi is the bytecode pointer
-  if (save_flags) {
-    __ mov(flags, rsi);
-    __ restore_bcp();
-  }
-}
-
-
-void TemplateTable::invokevirtual_helper(Register index,
-                                         Register recv,
-                                         Register flags) {
-  // Uses temporary registers rax, rdx
-  assert_different_registers(index, recv, rax, rdx);
-  assert(index == rbx, "");
-  assert(recv  == rcx, "");
-
-  // Test for an invoke of a final method
-  Label notFinal;
-  __ movl(rax, flags);
-  __ andl(rax, (1 << ConstantPoolCacheEntry::is_vfinal_shift));
-  __ jcc(Assembler::zero, notFinal);
-
-  const Register method = index;  // method must be rbx
-  assert(method == rbx,
-         "Method* must be rbx for interpreter calling convention");
-
-  // do the call - the index is actually the method to call
-  // that is, f2 is a vtable index if !is_vfinal, else f2 is a Method*
-
-  // It's final, need a null check here!
-  __ null_check(recv);
-
-  // profile this call
-  __ profile_final_call(rax);
-  __ profile_arguments_type(rax, method, rsi, true);
-
-  __ jump_from_interpreted(method, rax);
-
-  __ bind(notFinal);
-
-  // get receiver klass
-  __ null_check(recv, oopDesc::klass_offset_in_bytes());
-  __ load_klass(rax, recv);
-
-  // profile this call
-  __ profile_virtual_call(rax, rdi, rdx);
-
-  // get target Method* & entry point
-  __ lookup_virtual_method(rax, index, method);
-  __ profile_arguments_type(rdx, method, rsi, true);
-  __ jump_from_interpreted(method, rdx);
-}
-
-
-void TemplateTable::invokevirtual(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f2_byte, "use this argument");
-  prepare_invoke(byte_no,
-                 rbx,    // method or vtable index
-                 noreg,  // unused itable index
-                 rcx, rdx); // recv, flags
-
-  // rbx: index
-  // rcx: receiver
-  // rdx: flags
-
-  invokevirtual_helper(rbx, rcx, rdx);
-}
-
-
-void TemplateTable::invokespecial(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rbx, noreg,  // get f1 Method*
-                 rcx);  // get receiver also for null check
-  __ verify_oop(rcx);
-  __ null_check(rcx);
-  // do the call
-  __ profile_call(rax);
-  __ profile_arguments_type(rax, rbx, rsi, false);
-  __ jump_from_interpreted(rbx, rax);
-}
-
-
-void TemplateTable::invokestatic(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rbx);  // get f1 Method*
-  // do the call
-  __ profile_call(rax);
-  __ profile_arguments_type(rax, rbx, rsi, false);
-  __ jump_from_interpreted(rbx, rax);
-}
-
-
-void TemplateTable::fast_invokevfinal(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f2_byte, "use this argument");
-  __ stop("fast_invokevfinal not used on x86");
-}
-
-
-void TemplateTable::invokeinterface(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rax, rbx,  // get f1 Klass*, f2 itable index
-                 rcx, rdx); // recv, flags
-
-  // rax: interface klass (from f1)
-  // rbx: itable index (from f2)
-  // rcx: receiver
-  // rdx: flags
-
-  // Special case of invokeinterface called for virtual method of
-  // java.lang.Object.  See cpCacheOop.cpp for details.
-  // This code isn't produced by javac, but could be produced by
-  // another compliant java compiler.
-  Label notMethod;
-  __ movl(rdi, rdx);
-  __ andl(rdi, (1 << ConstantPoolCacheEntry::is_forced_virtual_shift));
-  __ jcc(Assembler::zero, notMethod);
-
-  invokevirtual_helper(rbx, rcx, rdx);
-  __ bind(notMethod);
-
-  // Get receiver klass into rdx - also a null check
-  __ restore_locals();  // restore rdi
-  __ null_check(rcx, oopDesc::klass_offset_in_bytes());
-  __ load_klass(rdx, rcx);
-
-  // profile this call
-  __ profile_virtual_call(rdx, rsi, rdi);
-
-  Label no_such_interface, no_such_method;
-
-  __ lookup_interface_method(// inputs: rec. class, interface, itable index
-                             rdx, rax, rbx,
-                             // outputs: method, scan temp. reg
-                             rbx, rsi,
-                             no_such_interface);
-
-  // rbx: Method* to call
-  // rcx: receiver
-  // Check for abstract method error
-  // Note: This should be done more efficiently via a throw_abstract_method_error
-  //       interpreter entry point and a conditional jump to it in case of a null
-  //       method.
-  __ testptr(rbx, rbx);
-  __ jcc(Assembler::zero, no_such_method);
-
-  __ profile_arguments_type(rdx, rbx, rsi, true);
-
-  // do the call
-  // rcx: receiver
-  // rbx,: Method*
-  __ jump_from_interpreted(rbx, rdx);
-  __ should_not_reach_here();
-
-  // exception handling code follows...
-  // note: must restore interpreter registers to canonical
-  //       state for exception handling to work correctly!
-
-  __ bind(no_such_method);
-  // throw exception
-  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
-  __ restore_bcp();      // rsi must be correct for exception handler   (was destroyed)
-  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::throw_AbstractMethodError));
-  // the call_VM checks for exception, so we should never return here.
-  __ should_not_reach_here();
-
-  __ bind(no_such_interface);
-  // throw exception
-  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
-  __ restore_bcp();      // rsi must be correct for exception handler   (was destroyed)
-  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address,
-                   InterpreterRuntime::throw_IncompatibleClassChangeError));
-  // the call_VM checks for exception, so we should never return here.
-  __ should_not_reach_here();
-}
-
-void TemplateTable::invokehandle(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  const Register rbx_method = rbx;
-  const Register rax_mtype  = rax;
-  const Register rcx_recv   = rcx;
-  const Register rdx_flags  = rdx;
-
-  prepare_invoke(byte_no, rbx_method, rax_mtype, rcx_recv);
-  __ verify_method_ptr(rbx_method);
-  __ verify_oop(rcx_recv);
-  __ null_check(rcx_recv);
-
-  // rax: MethodType object (from cpool->resolved_references[f1], if necessary)
-  // rbx: MH.invokeExact_MT method (from f2)
-
-  // Note:  rax_mtype is already pushed (if necessary) by prepare_invoke
-
-  // FIXME: profile the LambdaForm also
-  __ profile_final_call(rax);
-  __ profile_arguments_type(rdx, rbx_method, rsi, true);
-
-  __ jump_from_interpreted(rbx_method, rdx);
-}
-
-
-void TemplateTable::invokedynamic(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-
-  const Register rbx_method   = rbx;
-  const Register rax_callsite = rax;
-
-  prepare_invoke(byte_no, rbx_method, rax_callsite);
-
-  // rax: CallSite object (from cpool->resolved_references[f1])
-  // rbx: MH.linkToCallSite method (from f2)
-
-  // Note:  rax_callsite is already pushed by prepare_invoke
-
-  // %%% should make a type profile for any invokedynamic that takes a ref argument
-  // profile this call
-  __ profile_call(rsi);
-  __ profile_arguments_type(rdx, rbx, rsi, false);
-
-  __ verify_oop(rax_callsite);
-
-  __ jump_from_interpreted(rbx_method, rdx);
-}
-
-//----------------------------------------------------------------------------------------------------
-// Allocation
-
-void TemplateTable::_new() {
-  transition(vtos, atos);
-  __ get_unsigned_2_byte_index_at_bcp(rdx, 1);
-  Label slow_case;
-  Label slow_case_no_pop;
-  Label done;
-  Label initialize_header;
-  Label initialize_object;  // including clearing the fields
-  Label allocate_shared;
-
-  __ get_cpool_and_tags(rcx, rax);
-
-  // Make sure the class we're about to instantiate has been resolved.
-  // This is done before loading InstanceKlass to be consistent with the order
-  // how Constant Pool is updated (see ConstantPool::klass_at_put)
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-  __ cmpb(Address(rax, rdx, Address::times_1, tags_offset), JVM_CONSTANT_Class);
-  __ jcc(Assembler::notEqual, slow_case_no_pop);
-
-  // get InstanceKlass
-  __ movptr(rcx, Address(rcx, rdx, Address::times_ptr, sizeof(ConstantPool)));
-  __ push(rcx);  // save the contexts of klass for initializing the header
-
-  // make sure klass is initialized & doesn't have finalizer
-  // make sure klass is fully initialized
-  __ cmpb(Address(rcx, InstanceKlass::init_state_offset()), InstanceKlass::fully_initialized);
-  __ jcc(Assembler::notEqual, slow_case);
-
-  // get instance_size in InstanceKlass (scaled to a count of bytes)
-  __ movl(rdx, Address(rcx, Klass::layout_helper_offset()));
-  // test to see if it has a finalizer or is malformed in some way
-  __ testl(rdx, Klass::_lh_instance_slow_path_bit);
-  __ jcc(Assembler::notZero, slow_case);
-
-  //
-  // Allocate the instance
-  // 1) Try to allocate in the TLAB
-  // 2) if fail and the object is large allocate in the shared Eden
-  // 3) if the above fails (or is not applicable), go to a slow case
-  // (creates a new TLAB, etc.)
-
-  const bool allow_shared_alloc =
-    Universe::heap()->supports_inline_contig_alloc();
-
-  const Register thread = rcx;
-  if (UseTLAB || allow_shared_alloc) {
-    __ get_thread(thread);
-  }
-
-  if (UseTLAB) {
-    __ movptr(rax, Address(thread, in_bytes(JavaThread::tlab_top_offset())));
-    __ lea(rbx, Address(rax, rdx, Address::times_1));
-    __ cmpptr(rbx, Address(thread, in_bytes(JavaThread::tlab_end_offset())));
-    __ jcc(Assembler::above, allow_shared_alloc ? allocate_shared : slow_case);
-    __ movptr(Address(thread, in_bytes(JavaThread::tlab_top_offset())), rbx);
-    if (ZeroTLAB) {
-      // the fields have been already cleared
-      __ jmp(initialize_header);
-    } else {
-      // initialize both the header and fields
-      __ jmp(initialize_object);
-    }
-  }
-
-  // Allocation in the shared Eden, if allowed.
-  //
-  // rdx: instance size in bytes
-  if (allow_shared_alloc) {
-    __ bind(allocate_shared);
-
-    ExternalAddress heap_top((address)Universe::heap()->top_addr());
-
-    Label retry;
-    __ bind(retry);
-    __ movptr(rax, heap_top);
-    __ lea(rbx, Address(rax, rdx, Address::times_1));
-    __ cmpptr(rbx, ExternalAddress((address)Universe::heap()->end_addr()));
-    __ jcc(Assembler::above, slow_case);
-
-    // Compare rax, with the top addr, and if still equal, store the new
-    // top addr in rbx, at the address of the top addr pointer. Sets ZF if was
-    // equal, and clears it otherwise. Use lock prefix for atomicity on MPs.
-    //
-    // rax,: object begin
-    // rbx,: object end
-    // rdx: instance size in bytes
-    __ locked_cmpxchgptr(rbx, heap_top);
-
-    // if someone beat us on the allocation, try again, otherwise continue
-    __ jcc(Assembler::notEqual, retry);
-
-    __ incr_allocated_bytes(thread, rdx, 0);
-  }
-
-  if (UseTLAB || Universe::heap()->supports_inline_contig_alloc()) {
-    // The object is initialized before the header.  If the object size is
-    // zero, go directly to the header initialization.
-    __ bind(initialize_object);
-    __ decrement(rdx, sizeof(oopDesc));
-    __ jcc(Assembler::zero, initialize_header);
-
-    // Initialize topmost object field, divide rdx by 8, check if odd and
-    // test if zero.
-    __ xorl(rcx, rcx);    // use zero reg to clear memory (shorter code)
-    __ shrl(rdx, LogBytesPerLong); // divide by 2*oopSize and set carry flag if odd
-
-    // rdx must have been multiple of 8
-#ifdef ASSERT
-    // make sure rdx was multiple of 8
-    Label L;
-    // Ignore partial flag stall after shrl() since it is debug VM
-    __ jccb(Assembler::carryClear, L);
-    __ stop("object size is not multiple of 2 - adjust this code");
-    __ bind(L);
-    // rdx must be > 0, no extra check needed here
-#endif
-
-    // initialize remaining object fields: rdx was a multiple of 8
-    { Label loop;
-    __ bind(loop);
-    __ movptr(Address(rax, rdx, Address::times_8, sizeof(oopDesc) - 1*oopSize), rcx);
-    NOT_LP64(__ movptr(Address(rax, rdx, Address::times_8, sizeof(oopDesc) - 2*oopSize), rcx));
-    __ decrement(rdx);
-    __ jcc(Assembler::notZero, loop);
-    }
-
-    // initialize object header only.
-    __ bind(initialize_header);
-    if (UseBiasedLocking) {
-      __ pop(rcx);   // get saved klass back in the register.
-      __ movptr(rbx, Address(rcx, Klass::prototype_header_offset()));
-      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes ()), rbx);
-    } else {
-      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes ()),
-                (int32_t)markOopDesc::prototype()); // header
-      __ pop(rcx);   // get saved klass back in the register.
-    }
-    __ store_klass(rax, rcx);  // klass
-
-    {
-      SkipIfEqual skip_if(_masm, &DTraceAllocProbes, 0);
-      // Trigger dtrace event for fastpath
-      __ push(atos);
-      __ call_VM_leaf(
-           CAST_FROM_FN_PTR(address, SharedRuntime::dtrace_object_alloc), rax);
-      __ pop(atos);
-    }
-
-    __ jmp(done);
-  }
-
-  // slow case
-  __ bind(slow_case);
-  __ pop(rcx);   // restore stack pointer to what it was when we came in.
-  __ bind(slow_case_no_pop);
-  __ get_constant_pool(rax);
-  __ get_unsigned_2_byte_index_at_bcp(rdx, 1);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::_new), rax, rdx);
-
-  // continue
-  __ bind(done);
-}
-
-
-void TemplateTable::newarray() {
-  transition(itos, atos);
-  __ push_i(rax);                                 // make sure everything is on the stack
-  __ load_unsigned_byte(rdx, at_bcp(1));
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::newarray), rdx, rax);
-  __ pop_i(rdx);                                  // discard size
-}
-
-
-void TemplateTable::anewarray() {
-  transition(itos, atos);
-  __ get_unsigned_2_byte_index_at_bcp(rdx, 1);
-  __ get_constant_pool(rcx);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::anewarray), rcx, rdx, rax);
-}
-
-
-void TemplateTable::arraylength() {
-  transition(atos, itos);
-  __ null_check(rax, arrayOopDesc::length_offset_in_bytes());
-  __ movl(rax, Address(rax, arrayOopDesc::length_offset_in_bytes()));
-}
-
-
-void TemplateTable::checkcast() {
-  transition(atos, atos);
-  Label done, is_null, ok_is_subtype, quicked, resolved;
-  __ testptr(rax, rax);   // Object is in EAX
-  __ jcc(Assembler::zero, is_null);
-
-  // Get cpool & tags index
-  __ get_cpool_and_tags(rcx, rdx); // ECX=cpool, EDX=tags array
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // EBX=index
-  // See if bytecode has already been quicked
-  __ cmpb(Address(rdx, rbx, Address::times_1, Array<u1>::base_offset_in_bytes()), JVM_CONSTANT_Class);
-  __ jcc(Assembler::equal, quicked);
-
-  __ push(atos);
-  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc) );
-  // vm_result_2 has metadata result
-  // borrow rdi from locals
-  __ get_thread(rdi);
-  __ get_vm_result_2(rax, rdi);
-  __ restore_locals();
-  __ pop_ptr(rdx);
-  __ jmpb(resolved);
-
-  // Get superklass in EAX and subklass in EBX
-  __ bind(quicked);
-  __ mov(rdx, rax);          // Save object in EDX; EAX needed for subtype check
-  __ movptr(rax, Address(rcx, rbx, Address::times_ptr, sizeof(ConstantPool)));
-
-  __ bind(resolved);
-  __ load_klass(rbx, rdx);
-
-  // Generate subtype check.  Blows ECX.  Resets EDI.  Object in EDX.
-  // Superklass in EAX.  Subklass in EBX.
-  __ gen_subtype_check( rbx, ok_is_subtype );
-
-  // Come here on failure
-  __ push(rdx);
-  // object is at TOS
-  __ jump(ExternalAddress(Interpreter::_throw_ClassCastException_entry));
-
-  // Come here on success
-  __ bind(ok_is_subtype);
-  __ mov(rax,rdx);           // Restore object in EDX
-
-  // Collect counts on whether this check-cast sees NULLs a lot or not.
-  if (ProfileInterpreter) {
-    __ jmp(done);
-    __ bind(is_null);
-    __ profile_null_seen(rcx);
-  } else {
-    __ bind(is_null);   // same as 'done'
-  }
-  __ bind(done);
-}
-
-
-void TemplateTable::instanceof() {
-  transition(atos, itos);
-  Label done, is_null, ok_is_subtype, quicked, resolved;
-  __ testptr(rax, rax);
-  __ jcc(Assembler::zero, is_null);
-
-  // Get cpool & tags index
-  __ get_cpool_and_tags(rcx, rdx); // ECX=cpool, EDX=tags array
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // EBX=index
-  // See if bytecode has already been quicked
-  __ cmpb(Address(rdx, rbx, Address::times_1, Array<u1>::base_offset_in_bytes()), JVM_CONSTANT_Class);
-  __ jcc(Assembler::equal, quicked);
-
-  __ push(atos);
-  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc) );
-  // vm_result_2 has metadata result
-  // borrow rdi from locals
-  __ get_thread(rdi);
-  __ get_vm_result_2(rax, rdi);
-  __ restore_locals();
-  __ pop_ptr(rdx);
-  __ load_klass(rdx, rdx);
-  __ jmp(resolved);
-
-  // Get superklass in EAX and subklass in EDX
-  __ bind(quicked);
-  __ load_klass(rdx, rax);
-  __ movptr(rax, Address(rcx, rbx, Address::times_ptr, sizeof(ConstantPool)));
-
-  __ bind(resolved);
-
-  // Generate subtype check.  Blows ECX.  Resets EDI.
-  // Superklass in EAX.  Subklass in EDX.
-  __ gen_subtype_check( rdx, ok_is_subtype );
-
-  // Come here on failure
-  __ xorl(rax,rax);
-  __ jmpb(done);
-  // Come here on success
-  __ bind(ok_is_subtype);
-  __ movl(rax, 1);
-
-  // Collect counts on whether this test sees NULLs a lot or not.
-  if (ProfileInterpreter) {
-    __ jmp(done);
-    __ bind(is_null);
-    __ profile_null_seen(rcx);
-  } else {
-    __ bind(is_null);   // same as 'done'
-  }
-  __ bind(done);
-  // rax, = 0: obj == NULL or  obj is not an instanceof the specified klass
-  // rax, = 1: obj != NULL and obj is     an instanceof the specified klass
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Breakpoints
-void TemplateTable::_breakpoint() {
-
-  // Note: We get here even if we are single stepping..
-  // jbug inists on setting breakpoints at every bytecode
-  // even if we are in single step mode.
-
-  transition(vtos, vtos);
-
-  // get the unpatched byte code
-  __ get_method(rcx);
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::get_original_bytecode_at), rcx, rsi);
-  __ mov(rbx, rax);
-
-  // post the breakpoint event
-  __ get_method(rcx);
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::_breakpoint), rcx, rsi);
-
-  // complete the execution of original bytecode
-  __ dispatch_only_normal(vtos);
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Exceptions
-
-void TemplateTable::athrow() {
-  transition(atos, vtos);
-  __ null_check(rax);
-  __ jump(ExternalAddress(Interpreter::throw_exception_entry()));
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Synchronization
-//
-// Note: monitorenter & exit are symmetric routines; which is reflected
-//       in the assembly code structure as well
-//
-// Stack layout:
-//
-// [expressions  ] <--- rsp               = expression stack top
-// ..
-// [expressions  ]
-// [monitor entry] <--- monitor block top = expression stack bot
-// ..
-// [monitor entry]
-// [frame data   ] <--- monitor block bot
-// ...
-// [saved rbp,    ] <--- rbp,
-
-
-void TemplateTable::monitorenter() {
-  transition(atos, vtos);
-
-  // check for NULL object
-  __ null_check(rax);
-
-  const Address monitor_block_top(rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
-  const Address monitor_block_bot(rbp, frame::interpreter_frame_initial_sp_offset        * wordSize);
-  const int entry_size =         (     frame::interpreter_frame_monitor_size()           * wordSize);
-  Label allocated;
-
-  // initialize entry pointer
-  __ xorl(rdx, rdx);                             // points to free slot or NULL
-
-  // find a free slot in the monitor block (result in rdx)
-  { Label entry, loop, exit;
-    __ movptr(rcx, monitor_block_top);           // points to current entry, starting with top-most entry
-
-    __ lea(rbx, monitor_block_bot);              // points to word before bottom of monitor block
-    __ jmpb(entry);
-
-    __ bind(loop);
-    __ cmpptr(Address(rcx, BasicObjectLock::obj_offset_in_bytes()), (int32_t)NULL_WORD);  // check if current entry is used
-    __ cmovptr(Assembler::equal, rdx, rcx);      // if not used then remember entry in rdx
-    __ cmpptr(rax, Address(rcx, BasicObjectLock::obj_offset_in_bytes()));   // check if current entry is for same object
-    __ jccb(Assembler::equal, exit);             // if same object then stop searching
-    __ addptr(rcx, entry_size);                  // otherwise advance to next entry
-    __ bind(entry);
-    __ cmpptr(rcx, rbx);                         // check if bottom reached
-    __ jcc(Assembler::notEqual, loop);           // if not at bottom then check this entry
-    __ bind(exit);
-  }
-
-  __ testptr(rdx, rdx);                          // check if a slot has been found
-  __ jccb(Assembler::notZero, allocated);        // if found, continue with that one
-
-  // allocate one if there's no free slot
-  { Label entry, loop;
-    // 1. compute new pointers                   // rsp: old expression stack top
-    __ movptr(rdx, monitor_block_bot);           // rdx: old expression stack bottom
-    __ subptr(rsp, entry_size);                  // move expression stack top
-    __ subptr(rdx, entry_size);                  // move expression stack bottom
-    __ mov(rcx, rsp);                            // set start value for copy loop
-    __ movptr(monitor_block_bot, rdx);           // set new monitor block top
-    __ jmp(entry);
-    // 2. move expression stack contents
-    __ bind(loop);
-    __ movptr(rbx, Address(rcx, entry_size));    // load expression stack word from old location
-    __ movptr(Address(rcx, 0), rbx);             // and store it at new location
-    __ addptr(rcx, wordSize);                    // advance to next word
-    __ bind(entry);
-    __ cmpptr(rcx, rdx);                         // check if bottom reached
-    __ jcc(Assembler::notEqual, loop);           // if not at bottom then copy next word
-  }
-
-  // call run-time routine
-  // rdx: points to monitor entry
-  __ bind(allocated);
-
-  // Increment bcp to point to the next bytecode, so exception handling for async. exceptions work correctly.
-  // The object has already been poped from the stack, so the expression stack looks correct.
-  __ increment(rsi);
-
-  __ movptr(Address(rdx, BasicObjectLock::obj_offset_in_bytes()), rax);     // store object
-  __ lock_object(rdx);
-
-  // check to make sure this monitor doesn't cause stack overflow after locking
-  __ save_bcp();  // in case of exception
-  __ generate_stack_overflow_check(0);
-
-  // The bcp has already been incremented. Just need to dispatch to next instruction.
-  __ dispatch_next(vtos);
-}
-
-
-void TemplateTable::monitorexit() {
-  transition(atos, vtos);
-
-  // check for NULL object
-  __ null_check(rax);
-
-  const Address monitor_block_top(rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
-  const Address monitor_block_bot(rbp, frame::interpreter_frame_initial_sp_offset        * wordSize);
-  const int entry_size =         (     frame::interpreter_frame_monitor_size()           * wordSize);
-  Label found;
-
-  // find matching slot
-  { Label entry, loop;
-    __ movptr(rdx, monitor_block_top);           // points to current entry, starting with top-most entry
-    __ lea(rbx, monitor_block_bot);             // points to word before bottom of monitor block
-    __ jmpb(entry);
-
-    __ bind(loop);
-    __ cmpptr(rax, Address(rdx, BasicObjectLock::obj_offset_in_bytes()));   // check if current entry is for same object
-    __ jcc(Assembler::equal, found);             // if same object then stop searching
-    __ addptr(rdx, entry_size);                  // otherwise advance to next entry
-    __ bind(entry);
-    __ cmpptr(rdx, rbx);                         // check if bottom reached
-    __ jcc(Assembler::notEqual, loop);           // if not at bottom then check this entry
-  }
-
-  // error handling. Unlocking was not block-structured
-  Label end;
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::throw_illegal_monitor_state_exception));
-  __ should_not_reach_here();
-
-  // call run-time routine
-  // rcx: points to monitor entry
-  __ bind(found);
-  __ push_ptr(rax);                                 // make sure object is on stack (contract with oopMaps)
-  __ unlock_object(rdx);
-  __ pop_ptr(rax);                                  // discard object
-  __ bind(end);
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Wide instructions
-
-void TemplateTable::wide() {
-  transition(vtos, vtos);
-  __ load_unsigned_byte(rbx, at_bcp(1));
-  ExternalAddress wtable((address)Interpreter::_wentry_point);
-  __ jump(ArrayAddress(wtable, Address(noreg, rbx, Address::times_ptr)));
-  // Note: the rsi increment step is part of the individual wide bytecode implementations
-}
-
-
-//----------------------------------------------------------------------------------------------------
-// Multi arrays
-
-void TemplateTable::multianewarray() {
-  transition(vtos, atos);
-  __ load_unsigned_byte(rax, at_bcp(3)); // get number of dimensions
-  // last dim is on top of stack; we want address of first one:
-  // first_addr = last_addr + (ndims - 1) * stackElementSize - 1*wordsize
-  // the latter wordSize to point to the beginning of the array.
-  __ lea(  rax, Address(rsp, rax, Interpreter::stackElementScale(), -wordSize));
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::multianewarray), rax);     // pass in rax,
-  __ load_unsigned_byte(rbx, at_bcp(3));
-  __ lea(rsp, Address(rsp, rbx, Interpreter::stackElementScale()));  // get rid of counts
-}
-
-#endif /* !CC_INTERP */
--- a/hotspot/src/cpu/x86/vm/templateTable_x86_64.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,3741 +0,0 @@
-/*
- * Copyright (c) 2003, 2014, 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 "asm/macroAssembler.hpp"
-#include "interpreter/interpreter.hpp"
-#include "interpreter/interpreterRuntime.hpp"
-#include "interpreter/interp_masm.hpp"
-#include "interpreter/templateTable.hpp"
-#include "memory/universe.inline.hpp"
-#include "oops/methodData.hpp"
-#include "oops/objArrayKlass.hpp"
-#include "oops/oop.inline.hpp"
-#include "prims/methodHandles.hpp"
-#include "runtime/sharedRuntime.hpp"
-#include "runtime/stubRoutines.hpp"
-#include "runtime/synchronizer.hpp"
-#include "utilities/macros.hpp"
-
-#ifndef CC_INTERP
-
-#define __ _masm->
-
-// Platform-dependent initialization
-
-void TemplateTable::pd_initialize() {
-  // No amd64 specific initialization
-}
-
-// Address computation: local variables
-
-static inline Address iaddress(int n) {
-  return Address(r14, Interpreter::local_offset_in_bytes(n));
-}
-
-static inline Address laddress(int n) {
-  return iaddress(n + 1);
-}
-
-static inline Address faddress(int n) {
-  return iaddress(n);
-}
-
-static inline Address daddress(int n) {
-  return laddress(n);
-}
-
-static inline Address aaddress(int n) {
-  return iaddress(n);
-}
-
-static inline Address iaddress(Register r) {
-  return Address(r14, r, Address::times_8);
-}
-
-static inline Address laddress(Register r) {
-  return Address(r14, r, Address::times_8, Interpreter::local_offset_in_bytes(1));
-}
-
-static inline Address faddress(Register r) {
-  return iaddress(r);
-}
-
-static inline Address daddress(Register r) {
-  return laddress(r);
-}
-
-static inline Address aaddress(Register r) {
-  return iaddress(r);
-}
-
-static inline Address at_rsp() {
-  return Address(rsp, 0);
-}
-
-// At top of Java expression stack which may be different than esp().  It
-// isn't for category 1 objects.
-static inline Address at_tos   () {
-  return Address(rsp,  Interpreter::expr_offset_in_bytes(0));
-}
-
-static inline Address at_tos_p1() {
-  return Address(rsp,  Interpreter::expr_offset_in_bytes(1));
-}
-
-static inline Address at_tos_p2() {
-  return Address(rsp,  Interpreter::expr_offset_in_bytes(2));
-}
-
-// Condition conversion
-static Assembler::Condition j_not(TemplateTable::Condition cc) {
-  switch (cc) {
-  case TemplateTable::equal        : return Assembler::notEqual;
-  case TemplateTable::not_equal    : return Assembler::equal;
-  case TemplateTable::less         : return Assembler::greaterEqual;
-  case TemplateTable::less_equal   : return Assembler::greater;
-  case TemplateTable::greater      : return Assembler::lessEqual;
-  case TemplateTable::greater_equal: return Assembler::less;
-  }
-  ShouldNotReachHere();
-  return Assembler::zero;
-}
-
-
-// Miscelaneous helper routines
-// Store an oop (or NULL) at the address described by obj.
-// If val == noreg this means store a NULL
-
-static void do_oop_store(InterpreterMacroAssembler* _masm,
-                         Address obj,
-                         Register val,
-                         BarrierSet::Name barrier,
-                         bool precise) {
-  assert(val == noreg || val == rax, "parameter is just for looks");
-  switch (barrier) {
-#if INCLUDE_ALL_GCS
-    case BarrierSet::G1SATBCT:
-    case BarrierSet::G1SATBCTLogging:
-      {
-        // flatten object address if needed
-        if (obj.index() == noreg && obj.disp() == 0) {
-          if (obj.base() != rdx) {
-            __ movq(rdx, obj.base());
-          }
-        } else {
-          __ leaq(rdx, obj);
-        }
-        __ g1_write_barrier_pre(rdx /* obj */,
-                                rbx /* pre_val */,
-                                r15_thread /* thread */,
-                                r8  /* tmp */,
-                                val != noreg /* tosca_live */,
-                                false /* expand_call */);
-        if (val == noreg) {
-          __ store_heap_oop_null(Address(rdx, 0));
-        } else {
-          // G1 barrier needs uncompressed oop for region cross check.
-          Register new_val = val;
-          if (UseCompressedOops) {
-            new_val = rbx;
-            __ movptr(new_val, val);
-          }
-          __ store_heap_oop(Address(rdx, 0), val);
-          __ g1_write_barrier_post(rdx /* store_adr */,
-                                   new_val /* new_val */,
-                                   r15_thread /* thread */,
-                                   r8 /* tmp */,
-                                   rbx /* tmp2 */);
-        }
-      }
-      break;
-#endif // INCLUDE_ALL_GCS
-    case BarrierSet::CardTableModRef:
-    case BarrierSet::CardTableExtension:
-      {
-        if (val == noreg) {
-          __ store_heap_oop_null(obj);
-        } else {
-          __ store_heap_oop(obj, val);
-          // flatten object address if needed
-          if (!precise || (obj.index() == noreg && obj.disp() == 0)) {
-            __ store_check(obj.base());
-          } else {
-            __ leaq(rdx, obj);
-            __ store_check(rdx);
-          }
-        }
-      }
-      break;
-    case BarrierSet::ModRef:
-      if (val == noreg) {
-        __ store_heap_oop_null(obj);
-      } else {
-        __ store_heap_oop(obj, val);
-      }
-      break;
-    default      :
-      ShouldNotReachHere();
-
-  }
-}
-
-Address TemplateTable::at_bcp(int offset) {
-  assert(_desc->uses_bcp(), "inconsistent uses_bcp information");
-  return Address(r13, offset);
-}
-
-void TemplateTable::patch_bytecode(Bytecodes::Code bc, Register bc_reg,
-                                   Register temp_reg, bool load_bc_into_bc_reg/*=true*/,
-                                   int byte_no) {
-  if (!RewriteBytecodes)  return;
-  Label L_patch_done;
-
-  switch (bc) {
-  case Bytecodes::_fast_aputfield:
-  case Bytecodes::_fast_bputfield:
-  case Bytecodes::_fast_cputfield:
-  case Bytecodes::_fast_dputfield:
-  case Bytecodes::_fast_fputfield:
-  case Bytecodes::_fast_iputfield:
-  case Bytecodes::_fast_lputfield:
-  case Bytecodes::_fast_sputfield:
-    {
-      // We skip bytecode quickening for putfield instructions when
-      // the put_code written to the constant pool cache is zero.
-      // This is required so that every execution of this instruction
-      // calls out to InterpreterRuntime::resolve_get_put to do
-      // additional, required work.
-      assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
-      assert(load_bc_into_bc_reg, "we use bc_reg as temp");
-      __ get_cache_and_index_and_bytecode_at_bcp(temp_reg, bc_reg, temp_reg, byte_no, 1);
-      __ movl(bc_reg, bc);
-      __ cmpl(temp_reg, (int) 0);
-      __ jcc(Assembler::zero, L_patch_done);  // don't patch
-    }
-    break;
-  default:
-    assert(byte_no == -1, "sanity");
-    // the pair bytecodes have already done the load.
-    if (load_bc_into_bc_reg) {
-      __ movl(bc_reg, bc);
-    }
-  }
-
-  if (JvmtiExport::can_post_breakpoint()) {
-    Label L_fast_patch;
-    // if a breakpoint is present we can't rewrite the stream directly
-    __ movzbl(temp_reg, at_bcp(0));
-    __ cmpl(temp_reg, Bytecodes::_breakpoint);
-    __ jcc(Assembler::notEqual, L_fast_patch);
-    __ get_method(temp_reg);
-    // Let breakpoint table handling rewrite to quicker bytecode
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::set_original_bytecode_at), temp_reg, r13, bc_reg);
-#ifndef ASSERT
-    __ jmpb(L_patch_done);
-#else
-    __ jmp(L_patch_done);
-#endif
-    __ bind(L_fast_patch);
-  }
-
-#ifdef ASSERT
-  Label L_okay;
-  __ load_unsigned_byte(temp_reg, at_bcp(0));
-  __ cmpl(temp_reg, (int) Bytecodes::java_code(bc));
-  __ jcc(Assembler::equal, L_okay);
-  __ cmpl(temp_reg, bc_reg);
-  __ jcc(Assembler::equal, L_okay);
-  __ stop("patching the wrong bytecode");
-  __ bind(L_okay);
-#endif
-
-  // patch bytecode
-  __ movb(at_bcp(0), bc_reg);
-  __ bind(L_patch_done);
-}
-
-
-// Individual instructions
-
-void TemplateTable::nop() {
-  transition(vtos, vtos);
-  // nothing to do
-}
-
-void TemplateTable::shouldnotreachhere() {
-  transition(vtos, vtos);
-  __ stop("shouldnotreachhere bytecode");
-}
-
-void TemplateTable::aconst_null() {
-  transition(vtos, atos);
-  __ xorl(rax, rax);
-}
-
-void TemplateTable::iconst(int value) {
-  transition(vtos, itos);
-  if (value == 0) {
-    __ xorl(rax, rax);
-  } else {
-    __ movl(rax, value);
-  }
-}
-
-void TemplateTable::lconst(int value) {
-  transition(vtos, ltos);
-  if (value == 0) {
-    __ xorl(rax, rax);
-  } else {
-    __ movl(rax, value);
-  }
-}
-
-void TemplateTable::fconst(int value) {
-  transition(vtos, ftos);
-  static float one = 1.0f, two = 2.0f;
-  switch (value) {
-  case 0:
-    __ xorps(xmm0, xmm0);
-    break;
-  case 1:
-    __ movflt(xmm0, ExternalAddress((address) &one));
-    break;
-  case 2:
-    __ movflt(xmm0, ExternalAddress((address) &two));
-    break;
-  default:
-    ShouldNotReachHere();
-    break;
-  }
-}
-
-void TemplateTable::dconst(int value) {
-  transition(vtos, dtos);
-  static double one = 1.0;
-  switch (value) {
-  case 0:
-    __ xorpd(xmm0, xmm0);
-    break;
-  case 1:
-    __ movdbl(xmm0, ExternalAddress((address) &one));
-    break;
-  default:
-    ShouldNotReachHere();
-    break;
-  }
-}
-
-void TemplateTable::bipush() {
-  transition(vtos, itos);
-  __ load_signed_byte(rax, at_bcp(1));
-}
-
-void TemplateTable::sipush() {
-  transition(vtos, itos);
-  __ load_unsigned_short(rax, at_bcp(1));
-  __ bswapl(rax);
-  __ sarl(rax, 16);
-}
-
-void TemplateTable::ldc(bool wide) {
-  transition(vtos, vtos);
-  Label call_ldc, notFloat, notClass, Done;
-
-  if (wide) {
-    __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
-  } else {
-    __ load_unsigned_byte(rbx, at_bcp(1));
-  }
-
-  __ get_cpool_and_tags(rcx, rax);
-  const int base_offset = ConstantPool::header_size() * wordSize;
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-
-  // get type
-  __ movzbl(rdx, Address(rax, rbx, Address::times_1, tags_offset));
-
-  // unresolved class - get the resolved class
-  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClass);
-  __ jccb(Assembler::equal, call_ldc);
-
-  // unresolved class in error state - call into runtime to throw the error
-  // from the first resolution attempt
-  __ cmpl(rdx, JVM_CONSTANT_UnresolvedClassInError);
-  __ jccb(Assembler::equal, call_ldc);
-
-  // resolved class - need to call vm to get java mirror of the class
-  __ cmpl(rdx, JVM_CONSTANT_Class);
-  __ jcc(Assembler::notEqual, notClass);
-
-  __ bind(call_ldc);
-  __ movl(c_rarg1, wide);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::ldc), c_rarg1);
-  __ push_ptr(rax);
-  __ verify_oop(rax);
-  __ jmp(Done);
-
-  __ bind(notClass);
-  __ cmpl(rdx, JVM_CONSTANT_Float);
-  __ jccb(Assembler::notEqual, notFloat);
-  // ftos
-  __ movflt(xmm0, Address(rcx, rbx, Address::times_8, base_offset));
-  __ push_f();
-  __ jmp(Done);
-
-  __ bind(notFloat);
-#ifdef ASSERT
-  {
-    Label L;
-    __ cmpl(rdx, JVM_CONSTANT_Integer);
-    __ jcc(Assembler::equal, L);
-    // String and Object are rewritten to fast_aldc
-    __ stop("unexpected tag type in ldc");
-    __ bind(L);
-  }
-#endif
-  // itos JVM_CONSTANT_Integer only
-  __ movl(rax, Address(rcx, rbx, Address::times_8, base_offset));
-  __ push_i(rax);
-  __ bind(Done);
-}
-
-// Fast path for caching oop constants.
-void TemplateTable::fast_aldc(bool wide) {
-  transition(vtos, atos);
-
-  Register result = rax;
-  Register tmp = rdx;
-  int index_size = wide ? sizeof(u2) : sizeof(u1);
-
-  Label resolved;
-
-  // We are resolved if the resolved reference cache entry contains a
-  // non-null object (String, MethodType, etc.)
-  assert_different_registers(result, tmp);
-  __ get_cache_index_at_bcp(tmp, 1, index_size);
-  __ load_resolved_reference_at_index(result, tmp);
-  __ testl(result, result);
-  __ jcc(Assembler::notZero, resolved);
-
-  address entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_ldc);
-
-  // first time invocation - must resolve first
-  __ movl(tmp, (int)bytecode());
-  __ call_VM(result, entry, tmp);
-
-  __ bind(resolved);
-
-  if (VerifyOops) {
-    __ verify_oop(result);
-  }
-}
-
-void TemplateTable::ldc2_w() {
-  transition(vtos, vtos);
-  Label Long, Done;
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1);
-
-  __ get_cpool_and_tags(rcx, rax);
-  const int base_offset = ConstantPool::header_size() * wordSize;
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-
-  // get type
-  __ cmpb(Address(rax, rbx, Address::times_1, tags_offset),
-          JVM_CONSTANT_Double);
-  __ jccb(Assembler::notEqual, Long);
-  // dtos
-  __ movdbl(xmm0, Address(rcx, rbx, Address::times_8, base_offset));
-  __ push_d();
-  __ jmpb(Done);
-
-  __ bind(Long);
-  // ltos
-  __ movq(rax, Address(rcx, rbx, Address::times_8, base_offset));
-  __ push_l();
-
-  __ bind(Done);
-}
-
-void TemplateTable::locals_index(Register reg, int offset) {
-  __ load_unsigned_byte(reg, at_bcp(offset));
-  __ negptr(reg);
-}
-
-void TemplateTable::iload() {
-  transition(vtos, itos);
-  if (RewriteFrequentPairs) {
-    Label rewrite, done;
-    const Register bc = c_rarg3;
-    assert(rbx != bc, "register damaged");
-
-    // get next byte
-    __ load_unsigned_byte(rbx,
-                          at_bcp(Bytecodes::length_for(Bytecodes::_iload)));
-    // if _iload, wait to rewrite to iload2.  We only want to rewrite the
-    // last two iloads in a pair.  Comparing against fast_iload means that
-    // the next bytecode is neither an iload or a caload, and therefore
-    // an iload pair.
-    __ cmpl(rbx, Bytecodes::_iload);
-    __ jcc(Assembler::equal, done);
-
-    __ cmpl(rbx, Bytecodes::_fast_iload);
-    __ movl(bc, Bytecodes::_fast_iload2);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _caload, rewrite to fast_icaload
-    __ cmpl(rbx, Bytecodes::_caload);
-    __ movl(bc, Bytecodes::_fast_icaload);
-    __ jccb(Assembler::equal, rewrite);
-
-    // rewrite so iload doesn't check again.
-    __ movl(bc, Bytecodes::_fast_iload);
-
-    // rewrite
-    // bc: fast bytecode
-    __ bind(rewrite);
-    patch_bytecode(Bytecodes::_iload, bc, rbx, false);
-    __ bind(done);
-  }
-
-  // Get the local value into tos
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-void TemplateTable::fast_iload2() {
-  transition(vtos, itos);
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-  __ push(itos);
-  locals_index(rbx, 3);
-  __ movl(rax, iaddress(rbx));
-}
-
-void TemplateTable::fast_iload() {
-  transition(vtos, itos);
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-void TemplateTable::lload() {
-  transition(vtos, ltos);
-  locals_index(rbx);
-  __ movq(rax, laddress(rbx));
-}
-
-void TemplateTable::fload() {
-  transition(vtos, ftos);
-  locals_index(rbx);
-  __ movflt(xmm0, faddress(rbx));
-}
-
-void TemplateTable::dload() {
-  transition(vtos, dtos);
-  locals_index(rbx);
-  __ movdbl(xmm0, daddress(rbx));
-}
-
-void TemplateTable::aload() {
-  transition(vtos, atos);
-  locals_index(rbx);
-  __ movptr(rax, aaddress(rbx));
-}
-
-void TemplateTable::locals_index_wide(Register reg) {
-  __ load_unsigned_short(reg, at_bcp(2));
-  __ bswapl(reg);
-  __ shrl(reg, 16);
-  __ negptr(reg);
-}
-
-void TemplateTable::wide_iload() {
-  transition(vtos, itos);
-  locals_index_wide(rbx);
-  __ movl(rax, iaddress(rbx));
-}
-
-void TemplateTable::wide_lload() {
-  transition(vtos, ltos);
-  locals_index_wide(rbx);
-  __ movq(rax, laddress(rbx));
-}
-
-void TemplateTable::wide_fload() {
-  transition(vtos, ftos);
-  locals_index_wide(rbx);
-  __ movflt(xmm0, faddress(rbx));
-}
-
-void TemplateTable::wide_dload() {
-  transition(vtos, dtos);
-  locals_index_wide(rbx);
-  __ movdbl(xmm0, daddress(rbx));
-}
-
-void TemplateTable::wide_aload() {
-  transition(vtos, atos);
-  locals_index_wide(rbx);
-  __ movptr(rax, aaddress(rbx));
-}
-
-void TemplateTable::index_check(Register array, Register index) {
-  // destroys rbx
-  // check array
-  __ null_check(array, arrayOopDesc::length_offset_in_bytes());
-  // sign extend index for use by indexed load
-  __ movl2ptr(index, index);
-  // check index
-  __ cmpl(index, Address(array, arrayOopDesc::length_offset_in_bytes()));
-  if (index != rbx) {
-    // ??? convention: move aberrant index into ebx for exception message
-    assert(rbx != array, "different registers");
-    __ movl(rbx, index);
-  }
-  __ jump_cc(Assembler::aboveEqual,
-             ExternalAddress(Interpreter::_throw_ArrayIndexOutOfBoundsException_entry));
-}
-
-void TemplateTable::iaload() {
-  transition(itos, itos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ movl(rax, Address(rdx, rax,
-                       Address::times_4,
-                       arrayOopDesc::base_offset_in_bytes(T_INT)));
-}
-
-void TemplateTable::laload() {
-  transition(itos, ltos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ movq(rax, Address(rdx, rbx,
-                       Address::times_8,
-                       arrayOopDesc::base_offset_in_bytes(T_LONG)));
-}
-
-void TemplateTable::faload() {
-  transition(itos, ftos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ movflt(xmm0, Address(rdx, rax,
-                         Address::times_4,
-                         arrayOopDesc::base_offset_in_bytes(T_FLOAT)));
-}
-
-void TemplateTable::daload() {
-  transition(itos, dtos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ movdbl(xmm0, Address(rdx, rax,
-                          Address::times_8,
-                          arrayOopDesc::base_offset_in_bytes(T_DOUBLE)));
-}
-
-void TemplateTable::aaload() {
-  transition(itos, atos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ load_heap_oop(rax, Address(rdx, rax,
-                                UseCompressedOops ? Address::times_4 : Address::times_8,
-                                arrayOopDesc::base_offset_in_bytes(T_OBJECT)));
-}
-
-void TemplateTable::baload() {
-  transition(itos, itos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ load_signed_byte(rax,
-                      Address(rdx, rax,
-                              Address::times_1,
-                              arrayOopDesc::base_offset_in_bytes(T_BYTE)));
-}
-
-void TemplateTable::caload() {
-  transition(itos, itos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ load_unsigned_short(rax,
-                         Address(rdx, rax,
-                                 Address::times_2,
-                                 arrayOopDesc::base_offset_in_bytes(T_CHAR)));
-}
-
-// iload followed by caload frequent pair
-void TemplateTable::fast_icaload() {
-  transition(vtos, itos);
-  // load index out of locals
-  locals_index(rbx);
-  __ movl(rax, iaddress(rbx));
-
-  // eax: index
-  // rdx: array
-  __ pop_ptr(rdx);
-  index_check(rdx, rax); // kills rbx
-  __ load_unsigned_short(rax,
-                         Address(rdx, rax,
-                                 Address::times_2,
-                                 arrayOopDesc::base_offset_in_bytes(T_CHAR)));
-}
-
-void TemplateTable::saload() {
-  transition(itos, itos);
-  __ pop_ptr(rdx);
-  // eax: index
-  // rdx: array
-  index_check(rdx, rax); // kills rbx
-  __ load_signed_short(rax,
-                       Address(rdx, rax,
-                               Address::times_2,
-                               arrayOopDesc::base_offset_in_bytes(T_SHORT)));
-}
-
-void TemplateTable::iload(int n) {
-  transition(vtos, itos);
-  __ movl(rax, iaddress(n));
-}
-
-void TemplateTable::lload(int n) {
-  transition(vtos, ltos);
-  __ movq(rax, laddress(n));
-}
-
-void TemplateTable::fload(int n) {
-  transition(vtos, ftos);
-  __ movflt(xmm0, faddress(n));
-}
-
-void TemplateTable::dload(int n) {
-  transition(vtos, dtos);
-  __ movdbl(xmm0, daddress(n));
-}
-
-void TemplateTable::aload(int n) {
-  transition(vtos, atos);
-  __ movptr(rax, aaddress(n));
-}
-
-void TemplateTable::aload_0() {
-  transition(vtos, atos);
-  // According to bytecode histograms, the pairs:
-  //
-  // _aload_0, _fast_igetfield
-  // _aload_0, _fast_agetfield
-  // _aload_0, _fast_fgetfield
-  //
-  // occur frequently. If RewriteFrequentPairs is set, the (slow)
-  // _aload_0 bytecode checks if the next bytecode is either
-  // _fast_igetfield, _fast_agetfield or _fast_fgetfield and then
-  // rewrites the current bytecode into a pair bytecode; otherwise it
-  // rewrites the current bytecode into _fast_aload_0 that doesn't do
-  // the pair check anymore.
-  //
-  // Note: If the next bytecode is _getfield, the rewrite must be
-  //       delayed, otherwise we may miss an opportunity for a pair.
-  //
-  // Also rewrite frequent pairs
-  //   aload_0, aload_1
-  //   aload_0, iload_1
-  // These bytecodes with a small amount of code are most profitable
-  // to rewrite
-  if (RewriteFrequentPairs) {
-    Label rewrite, done;
-    const Register bc = c_rarg3;
-    assert(rbx != bc, "register damaged");
-    // get next byte
-    __ load_unsigned_byte(rbx,
-                          at_bcp(Bytecodes::length_for(Bytecodes::_aload_0)));
-
-    // do actual aload_0
-    aload(0);
-
-    // if _getfield then wait with rewrite
-    __ cmpl(rbx, Bytecodes::_getfield);
-    __ jcc(Assembler::equal, done);
-
-    // if _igetfield then reqrite to _fast_iaccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_iaccess_0) ==
-           Bytecodes::_aload_0,
-           "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_igetfield);
-    __ movl(bc, Bytecodes::_fast_iaccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _agetfield then reqrite to _fast_aaccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_aaccess_0) ==
-           Bytecodes::_aload_0,
-           "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_agetfield);
-    __ movl(bc, Bytecodes::_fast_aaccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // if _fgetfield then reqrite to _fast_faccess_0
-    assert(Bytecodes::java_code(Bytecodes::_fast_faccess_0) ==
-           Bytecodes::_aload_0,
-           "fix bytecode definition");
-    __ cmpl(rbx, Bytecodes::_fast_fgetfield);
-    __ movl(bc, Bytecodes::_fast_faccess_0);
-    __ jccb(Assembler::equal, rewrite);
-
-    // else rewrite to _fast_aload0
-    assert(Bytecodes::java_code(Bytecodes::_fast_aload_0) ==
-           Bytecodes::_aload_0,
-           "fix bytecode definition");
-    __ movl(bc, Bytecodes::_fast_aload_0);
-
-    // rewrite
-    // bc: fast bytecode
-    __ bind(rewrite);
-    patch_bytecode(Bytecodes::_aload_0, bc, rbx, false);
-
-    __ bind(done);
-  } else {
-    aload(0);
-  }
-}
-
-void TemplateTable::istore() {
-  transition(itos, vtos);
-  locals_index(rbx);
-  __ movl(iaddress(rbx), rax);
-}
-
-void TemplateTable::lstore() {
-  transition(ltos, vtos);
-  locals_index(rbx);
-  __ movq(laddress(rbx), rax);
-}
-
-void TemplateTable::fstore() {
-  transition(ftos, vtos);
-  locals_index(rbx);
-  __ movflt(faddress(rbx), xmm0);
-}
-
-void TemplateTable::dstore() {
-  transition(dtos, vtos);
-  locals_index(rbx);
-  __ movdbl(daddress(rbx), xmm0);
-}
-
-void TemplateTable::astore() {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  locals_index(rbx);
-  __ movptr(aaddress(rbx), rax);
-}
-
-void TemplateTable::wide_istore() {
-  transition(vtos, vtos);
-  __ pop_i();
-  locals_index_wide(rbx);
-  __ movl(iaddress(rbx), rax);
-}
-
-void TemplateTable::wide_lstore() {
-  transition(vtos, vtos);
-  __ pop_l();
-  locals_index_wide(rbx);
-  __ movq(laddress(rbx), rax);
-}
-
-void TemplateTable::wide_fstore() {
-  transition(vtos, vtos);
-  __ pop_f();
-  locals_index_wide(rbx);
-  __ movflt(faddress(rbx), xmm0);
-}
-
-void TemplateTable::wide_dstore() {
-  transition(vtos, vtos);
-  __ pop_d();
-  locals_index_wide(rbx);
-  __ movdbl(daddress(rbx), xmm0);
-}
-
-void TemplateTable::wide_astore() {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  locals_index_wide(rbx);
-  __ movptr(aaddress(rbx), rax);
-}
-
-void TemplateTable::iastore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // eax: value
-  // ebx: index
-  // rdx: array
-  index_check(rdx, rbx); // prefer index in ebx
-  __ movl(Address(rdx, rbx,
-                  Address::times_4,
-                  arrayOopDesc::base_offset_in_bytes(T_INT)),
-          rax);
-}
-
-void TemplateTable::lastore() {
-  transition(ltos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // rax: value
-  // ebx: index
-  // rdx: array
-  index_check(rdx, rbx); // prefer index in ebx
-  __ movq(Address(rdx, rbx,
-                  Address::times_8,
-                  arrayOopDesc::base_offset_in_bytes(T_LONG)),
-          rax);
-}
-
-void TemplateTable::fastore() {
-  transition(ftos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // xmm0: value
-  // ebx:  index
-  // rdx:  array
-  index_check(rdx, rbx); // prefer index in ebx
-  __ movflt(Address(rdx, rbx,
-                   Address::times_4,
-                   arrayOopDesc::base_offset_in_bytes(T_FLOAT)),
-           xmm0);
-}
-
-void TemplateTable::dastore() {
-  transition(dtos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // xmm0: value
-  // ebx:  index
-  // rdx:  array
-  index_check(rdx, rbx); // prefer index in ebx
-  __ movdbl(Address(rdx, rbx,
-                   Address::times_8,
-                   arrayOopDesc::base_offset_in_bytes(T_DOUBLE)),
-           xmm0);
-}
-
-void TemplateTable::aastore() {
-  Label is_null, ok_is_subtype, done;
-  transition(vtos, vtos);
-  // stack: ..., array, index, value
-  __ movptr(rax, at_tos());    // value
-  __ movl(rcx, at_tos_p1()); // index
-  __ movptr(rdx, at_tos_p2()); // array
-
-  Address element_address(rdx, rcx,
-                          UseCompressedOops? Address::times_4 : Address::times_8,
-                          arrayOopDesc::base_offset_in_bytes(T_OBJECT));
-
-  index_check(rdx, rcx);     // kills rbx
-  // do array store check - check for NULL value first
-  __ testptr(rax, rax);
-  __ jcc(Assembler::zero, is_null);
-
-  // Move subklass into rbx
-  __ load_klass(rbx, rax);
-  // Move superklass into rax
-  __ load_klass(rax, rdx);
-  __ movptr(rax, Address(rax,
-                         ObjArrayKlass::element_klass_offset()));
-  // Compress array + index*oopSize + 12 into a single register.  Frees rcx.
-  __ lea(rdx, element_address);
-
-  // Generate subtype check.  Blows rcx, rdi
-  // Superklass in rax.  Subklass in rbx.
-  __ gen_subtype_check(rbx, ok_is_subtype);
-
-  // Come here on failure
-  // object is at TOS
-  __ jump(ExternalAddress(Interpreter::_throw_ArrayStoreException_entry));
-
-  // Come here on success
-  __ bind(ok_is_subtype);
-
-  // Get the value we will store
-  __ movptr(rax, at_tos());
-  // Now store using the appropriate barrier
-  do_oop_store(_masm, Address(rdx, 0), rax, _bs->kind(), true);
-  __ jmp(done);
-
-  // Have a NULL in rax, rdx=array, ecx=index.  Store NULL at ary[idx]
-  __ bind(is_null);
-  __ profile_null_seen(rbx);
-
-  // Store a NULL
-  do_oop_store(_masm, element_address, noreg, _bs->kind(), true);
-
-  // Pop stack arguments
-  __ bind(done);
-  __ addptr(rsp, 3 * Interpreter::stackElementSize);
-}
-
-void TemplateTable::bastore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // eax: value
-  // ebx: index
-  // rdx: array
-  index_check(rdx, rbx); // prefer index in ebx
-  __ movb(Address(rdx, rbx,
-                  Address::times_1,
-                  arrayOopDesc::base_offset_in_bytes(T_BYTE)),
-          rax);
-}
-
-void TemplateTable::castore() {
-  transition(itos, vtos);
-  __ pop_i(rbx);
-  __ pop_ptr(rdx);
-  // eax: value
-  // ebx: index
-  // rdx: array
-  index_check(rdx, rbx);  // prefer index in ebx
-  __ movw(Address(rdx, rbx,
-                  Address::times_2,
-                  arrayOopDesc::base_offset_in_bytes(T_CHAR)),
-          rax);
-}
-
-void TemplateTable::sastore() {
-  castore();
-}
-
-void TemplateTable::istore(int n) {
-  transition(itos, vtos);
-  __ movl(iaddress(n), rax);
-}
-
-void TemplateTable::lstore(int n) {
-  transition(ltos, vtos);
-  __ movq(laddress(n), rax);
-}
-
-void TemplateTable::fstore(int n) {
-  transition(ftos, vtos);
-  __ movflt(faddress(n), xmm0);
-}
-
-void TemplateTable::dstore(int n) {
-  transition(dtos, vtos);
-  __ movdbl(daddress(n), xmm0);
-}
-
-void TemplateTable::astore(int n) {
-  transition(vtos, vtos);
-  __ pop_ptr(rax);
-  __ movptr(aaddress(n), rax);
-}
-
-void TemplateTable::pop() {
-  transition(vtos, vtos);
-  __ addptr(rsp, Interpreter::stackElementSize);
-}
-
-void TemplateTable::pop2() {
-  transition(vtos, vtos);
-  __ addptr(rsp, 2 * Interpreter::stackElementSize);
-}
-
-void TemplateTable::dup() {
-  transition(vtos, vtos);
-  __ load_ptr(0, rax);
-  __ push_ptr(rax);
-  // stack: ..., a, a
-}
-
-void TemplateTable::dup_x1() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr( 0, rax);  // load b
-  __ load_ptr( 1, rcx);  // load a
-  __ store_ptr(1, rax);  // store b
-  __ store_ptr(0, rcx);  // store a
-  __ push_ptr(rax);      // push b
-  // stack: ..., b, a, b
-}
-
-void TemplateTable::dup_x2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c
-  __ load_ptr( 0, rax);  // load c
-  __ load_ptr( 2, rcx);  // load a
-  __ store_ptr(2, rax);  // store c in a
-  __ push_ptr(rax);      // push c
-  // stack: ..., c, b, c, c
-  __ load_ptr( 2, rax);  // load b
-  __ store_ptr(2, rcx);  // store a in b
-  // stack: ..., c, a, c, c
-  __ store_ptr(1, rax);  // store b in c
-  // stack: ..., c, a, b, c
-}
-
-void TemplateTable::dup2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr(1, rax);  // load a
-  __ push_ptr(rax);     // push a
-  __ load_ptr(1, rax);  // load b
-  __ push_ptr(rax);     // push b
-  // stack: ..., a, b, a, b
-}
-
-void TemplateTable::dup2_x1() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c
-  __ load_ptr( 0, rcx);  // load c
-  __ load_ptr( 1, rax);  // load b
-  __ push_ptr(rax);      // push b
-  __ push_ptr(rcx);      // push c
-  // stack: ..., a, b, c, b, c
-  __ store_ptr(3, rcx);  // store c in b
-  // stack: ..., a, c, c, b, c
-  __ load_ptr( 4, rcx);  // load a
-  __ store_ptr(2, rcx);  // store a in 2nd c
-  // stack: ..., a, c, a, b, c
-  __ store_ptr(4, rax);  // store b in a
-  // stack: ..., b, c, a, b, c
-}
-
-void TemplateTable::dup2_x2() {
-  transition(vtos, vtos);
-  // stack: ..., a, b, c, d
-  __ load_ptr( 0, rcx);  // load d
-  __ load_ptr( 1, rax);  // load c
-  __ push_ptr(rax);      // push c
-  __ push_ptr(rcx);      // push d
-  // stack: ..., a, b, c, d, c, d
-  __ load_ptr( 4, rax);  // load b
-  __ store_ptr(2, rax);  // store b in d
-  __ store_ptr(4, rcx);  // store d in b
-  // stack: ..., a, d, c, b, c, d
-  __ load_ptr( 5, rcx);  // load a
-  __ load_ptr( 3, rax);  // load c
-  __ store_ptr(3, rcx);  // store a in c
-  __ store_ptr(5, rax);  // store c in a
-  // stack: ..., c, d, a, b, c, d
-}
-
-void TemplateTable::swap() {
-  transition(vtos, vtos);
-  // stack: ..., a, b
-  __ load_ptr( 1, rcx);  // load a
-  __ load_ptr( 0, rax);  // load b
-  __ store_ptr(0, rcx);  // store a in b
-  __ store_ptr(1, rax);  // store b in a
-  // stack: ..., b, a
-}
-
-void TemplateTable::iop2(Operation op) {
-  transition(itos, itos);
-  switch (op) {
-  case add  :                    __ pop_i(rdx); __ addl (rax, rdx); break;
-  case sub  : __ movl(rdx, rax); __ pop_i(rax); __ subl (rax, rdx); break;
-  case mul  :                    __ pop_i(rdx); __ imull(rax, rdx); break;
-  case _and :                    __ pop_i(rdx); __ andl (rax, rdx); break;
-  case _or  :                    __ pop_i(rdx); __ orl  (rax, rdx); break;
-  case _xor :                    __ pop_i(rdx); __ xorl (rax, rdx); break;
-  case shl  : __ movl(rcx, rax); __ pop_i(rax); __ shll (rax);      break;
-  case shr  : __ movl(rcx, rax); __ pop_i(rax); __ sarl (rax);      break;
-  case ushr : __ movl(rcx, rax); __ pop_i(rax); __ shrl (rax);      break;
-  default   : ShouldNotReachHere();
-  }
-}
-
-void TemplateTable::lop2(Operation op) {
-  transition(ltos, ltos);
-  switch (op) {
-  case add  :                    __ pop_l(rdx); __ addptr(rax, rdx); break;
-  case sub  : __ mov(rdx, rax);  __ pop_l(rax); __ subptr(rax, rdx); break;
-  case _and :                    __ pop_l(rdx); __ andptr(rax, rdx); break;
-  case _or  :                    __ pop_l(rdx); __ orptr (rax, rdx); break;
-  case _xor :                    __ pop_l(rdx); __ xorptr(rax, rdx); break;
-  default   : ShouldNotReachHere();
-  }
-}
-
-void TemplateTable::idiv() {
-  transition(itos, itos);
-  __ movl(rcx, rax);
-  __ pop_i(rax);
-  // Note: could xor eax and ecx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivl(rcx);
-}
-
-void TemplateTable::irem() {
-  transition(itos, itos);
-  __ movl(rcx, rax);
-  __ pop_i(rax);
-  // Note: could xor eax and ecx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivl(rcx);
-  __ movl(rax, rdx);
-}
-
-void TemplateTable::lmul() {
-  transition(ltos, ltos);
-  __ pop_l(rdx);
-  __ imulq(rax, rdx);
-}
-
-void TemplateTable::ldiv() {
-  transition(ltos, ltos);
-  __ mov(rcx, rax);
-  __ pop_l(rax);
-  // generate explicit div0 check
-  __ testq(rcx, rcx);
-  __ jump_cc(Assembler::zero,
-             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
-  // Note: could xor rax and rcx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivq(rcx); // kills rbx
-}
-
-void TemplateTable::lrem() {
-  transition(ltos, ltos);
-  __ mov(rcx, rax);
-  __ pop_l(rax);
-  __ testq(rcx, rcx);
-  __ jump_cc(Assembler::zero,
-             ExternalAddress(Interpreter::_throw_ArithmeticException_entry));
-  // Note: could xor rax and rcx and compare with (-1 ^ min_int). If
-  //       they are not equal, one could do a normal division (no correction
-  //       needed), which may speed up this implementation for the common case.
-  //       (see also JVM spec., p.243 & p.271)
-  __ corrected_idivq(rcx); // kills rbx
-  __ mov(rax, rdx);
-}
-
-void TemplateTable::lshl() {
-  transition(itos, ltos);
-  __ movl(rcx, rax);                             // get shift count
-  __ pop_l(rax);                                 // get shift value
-  __ shlq(rax);
-}
-
-void TemplateTable::lshr() {
-  transition(itos, ltos);
-  __ movl(rcx, rax);                             // get shift count
-  __ pop_l(rax);                                 // get shift value
-  __ sarq(rax);
-}
-
-void TemplateTable::lushr() {
-  transition(itos, ltos);
-  __ movl(rcx, rax);                             // get shift count
-  __ pop_l(rax);                                 // get shift value
-  __ shrq(rax);
-}
-
-void TemplateTable::fop2(Operation op) {
-  transition(ftos, ftos);
-  switch (op) {
-  case add:
-    __ addss(xmm0, at_rsp());
-    __ addptr(rsp, Interpreter::stackElementSize);
-    break;
-  case sub:
-    __ movflt(xmm1, xmm0);
-    __ pop_f(xmm0);
-    __ subss(xmm0, xmm1);
-    break;
-  case mul:
-    __ mulss(xmm0, at_rsp());
-    __ addptr(rsp, Interpreter::stackElementSize);
-    break;
-  case div:
-    __ movflt(xmm1, xmm0);
-    __ pop_f(xmm0);
-    __ divss(xmm0, xmm1);
-    break;
-  case rem:
-    __ movflt(xmm1, xmm0);
-    __ pop_f(xmm0);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::frem), 2);
-    break;
-  default:
-    ShouldNotReachHere();
-    break;
-  }
-}
-
-void TemplateTable::dop2(Operation op) {
-  transition(dtos, dtos);
-  switch (op) {
-  case add:
-    __ addsd(xmm0, at_rsp());
-    __ addptr(rsp, 2 * Interpreter::stackElementSize);
-    break;
-  case sub:
-    __ movdbl(xmm1, xmm0);
-    __ pop_d(xmm0);
-    __ subsd(xmm0, xmm1);
-    break;
-  case mul:
-    __ mulsd(xmm0, at_rsp());
-    __ addptr(rsp, 2 * Interpreter::stackElementSize);
-    break;
-  case div:
-    __ movdbl(xmm1, xmm0);
-    __ pop_d(xmm0);
-    __ divsd(xmm0, xmm1);
-    break;
-  case rem:
-    __ movdbl(xmm1, xmm0);
-    __ pop_d(xmm0);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::drem), 2);
-    break;
-  default:
-    ShouldNotReachHere();
-    break;
-  }
-}
-
-void TemplateTable::ineg() {
-  transition(itos, itos);
-  __ negl(rax);
-}
-
-void TemplateTable::lneg() {
-  transition(ltos, ltos);
-  __ negq(rax);
-}
-
-// Note: 'double' and 'long long' have 32-bits alignment on x86.
-static jlong* double_quadword(jlong *adr, jlong lo, jlong hi) {
-  // Use the expression (adr)&(~0xF) to provide 128-bits aligned address
-  // of 128-bits operands for SSE instructions.
-  jlong *operand = (jlong*)(((intptr_t)adr)&((intptr_t)(~0xF)));
-  // Store the value to a 128-bits operand.
-  operand[0] = lo;
-  operand[1] = hi;
-  return operand;
-}
-
-// Buffer for 128-bits masks used by SSE instructions.
-static jlong float_signflip_pool[2*2];
-static jlong double_signflip_pool[2*2];
-
-void TemplateTable::fneg() {
-  transition(ftos, ftos);
-  static jlong *float_signflip  = double_quadword(&float_signflip_pool[1], 0x8000000080000000, 0x8000000080000000);
-  __ xorps(xmm0, ExternalAddress((address) float_signflip));
-}
-
-void TemplateTable::dneg() {
-  transition(dtos, dtos);
-  static jlong *double_signflip  = double_quadword(&double_signflip_pool[1], 0x8000000000000000, 0x8000000000000000);
-  __ xorpd(xmm0, ExternalAddress((address) double_signflip));
-}
-
-void TemplateTable::iinc() {
-  transition(vtos, vtos);
-  __ load_signed_byte(rdx, at_bcp(2)); // get constant
-  locals_index(rbx);
-  __ addl(iaddress(rbx), rdx);
-}
-
-void TemplateTable::wide_iinc() {
-  transition(vtos, vtos);
-  __ movl(rdx, at_bcp(4)); // get constant
-  locals_index_wide(rbx);
-  __ bswapl(rdx); // swap bytes & sign-extend constant
-  __ sarl(rdx, 16);
-  __ addl(iaddress(rbx), rdx);
-  // Note: should probably use only one movl to get both
-  //       the index and the constant -> fix this
-}
-
-void TemplateTable::convert() {
-  // Checking
-#ifdef ASSERT
-  {
-    TosState tos_in  = ilgl;
-    TosState tos_out = ilgl;
-    switch (bytecode()) {
-    case Bytecodes::_i2l: // fall through
-    case Bytecodes::_i2f: // fall through
-    case Bytecodes::_i2d: // fall through
-    case Bytecodes::_i2b: // fall through
-    case Bytecodes::_i2c: // fall through
-    case Bytecodes::_i2s: tos_in = itos; break;
-    case Bytecodes::_l2i: // fall through
-    case Bytecodes::_l2f: // fall through
-    case Bytecodes::_l2d: tos_in = ltos; break;
-    case Bytecodes::_f2i: // fall through
-    case Bytecodes::_f2l: // fall through
-    case Bytecodes::_f2d: tos_in = ftos; break;
-    case Bytecodes::_d2i: // fall through
-    case Bytecodes::_d2l: // fall through
-    case Bytecodes::_d2f: tos_in = dtos; break;
-    default             : ShouldNotReachHere();
-    }
-    switch (bytecode()) {
-    case Bytecodes::_l2i: // fall through
-    case Bytecodes::_f2i: // fall through
-    case Bytecodes::_d2i: // fall through
-    case Bytecodes::_i2b: // fall through
-    case Bytecodes::_i2c: // fall through
-    case Bytecodes::_i2s: tos_out = itos; break;
-    case Bytecodes::_i2l: // fall through
-    case Bytecodes::_f2l: // fall through
-    case Bytecodes::_d2l: tos_out = ltos; break;
-    case Bytecodes::_i2f: // fall through
-    case Bytecodes::_l2f: // fall through
-    case Bytecodes::_d2f: tos_out = ftos; break;
-    case Bytecodes::_i2d: // fall through
-    case Bytecodes::_l2d: // fall through
-    case Bytecodes::_f2d: tos_out = dtos; break;
-    default             : ShouldNotReachHere();
-    }
-    transition(tos_in, tos_out);
-  }
-#endif // ASSERT
-
-  static const int64_t is_nan = 0x8000000000000000L;
-
-  // Conversion
-  switch (bytecode()) {
-  case Bytecodes::_i2l:
-    __ movslq(rax, rax);
-    break;
-  case Bytecodes::_i2f:
-    __ cvtsi2ssl(xmm0, rax);
-    break;
-  case Bytecodes::_i2d:
-    __ cvtsi2sdl(xmm0, rax);
-    break;
-  case Bytecodes::_i2b:
-    __ movsbl(rax, rax);
-    break;
-  case Bytecodes::_i2c:
-    __ movzwl(rax, rax);
-    break;
-  case Bytecodes::_i2s:
-    __ movswl(rax, rax);
-    break;
-  case Bytecodes::_l2i:
-    __ movl(rax, rax);
-    break;
-  case Bytecodes::_l2f:
-    __ cvtsi2ssq(xmm0, rax);
-    break;
-  case Bytecodes::_l2d:
-    __ cvtsi2sdq(xmm0, rax);
-    break;
-  case Bytecodes::_f2i:
-  {
-    Label L;
-    __ cvttss2sil(rax, xmm0);
-    __ cmpl(rax, 0x80000000); // NaN or overflow/underflow?
-    __ jcc(Assembler::notEqual, L);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2i), 1);
-    __ bind(L);
-  }
-    break;
-  case Bytecodes::_f2l:
-  {
-    Label L;
-    __ cvttss2siq(rax, xmm0);
-    // NaN or overflow/underflow?
-    __ cmp64(rax, ExternalAddress((address) &is_nan));
-    __ jcc(Assembler::notEqual, L);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::f2l), 1);
-    __ bind(L);
-  }
-    break;
-  case Bytecodes::_f2d:
-    __ cvtss2sd(xmm0, xmm0);
-    break;
-  case Bytecodes::_d2i:
-  {
-    Label L;
-    __ cvttsd2sil(rax, xmm0);
-    __ cmpl(rax, 0x80000000); // NaN or overflow/underflow?
-    __ jcc(Assembler::notEqual, L);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2i), 1);
-    __ bind(L);
-  }
-    break;
-  case Bytecodes::_d2l:
-  {
-    Label L;
-    __ cvttsd2siq(rax, xmm0);
-    // NaN or overflow/underflow?
-    __ cmp64(rax, ExternalAddress((address) &is_nan));
-    __ jcc(Assembler::notEqual, L);
-    __ call_VM_leaf(CAST_FROM_FN_PTR(address, SharedRuntime::d2l), 1);
-    __ bind(L);
-  }
-    break;
-  case Bytecodes::_d2f:
-    __ cvtsd2ss(xmm0, xmm0);
-    break;
-  default:
-    ShouldNotReachHere();
-  }
-}
-
-void TemplateTable::lcmp() {
-  transition(ltos, itos);
-  Label done;
-  __ pop_l(rdx);
-  __ cmpq(rdx, rax);
-  __ movl(rax, -1);
-  __ jccb(Assembler::less, done);
-  __ setb(Assembler::notEqual, rax);
-  __ movzbl(rax, rax);
-  __ bind(done);
-}
-
-void TemplateTable::float_cmp(bool is_float, int unordered_result) {
-  Label done;
-  if (is_float) {
-    // XXX get rid of pop here, use ... reg, mem32
-    __ pop_f(xmm1);
-    __ ucomiss(xmm1, xmm0);
-  } else {
-    // XXX get rid of pop here, use ... reg, mem64
-    __ pop_d(xmm1);
-    __ ucomisd(xmm1, xmm0);
-  }
-  if (unordered_result < 0) {
-    __ movl(rax, -1);
-    __ jccb(Assembler::parity, done);
-    __ jccb(Assembler::below, done);
-    __ setb(Assembler::notEqual, rdx);
-    __ movzbl(rax, rdx);
-  } else {
-    __ movl(rax, 1);
-    __ jccb(Assembler::parity, done);
-    __ jccb(Assembler::above, done);
-    __ movl(rax, 0);
-    __ jccb(Assembler::equal, done);
-    __ decrementl(rax);
-  }
-  __ bind(done);
-}
-
-void TemplateTable::branch(bool is_jsr, bool is_wide) {
-  __ get_method(rcx); // rcx holds method
-  __ profile_taken_branch(rax, rbx); // rax holds updated MDP, rbx
-                                     // holds bumped taken count
-
-  const ByteSize be_offset = MethodCounters::backedge_counter_offset() +
-                             InvocationCounter::counter_offset();
-  const ByteSize inv_offset = MethodCounters::invocation_counter_offset() +
-                              InvocationCounter::counter_offset();
-
-  // Load up edx with the branch displacement
-  if (is_wide) {
-    __ movl(rdx, at_bcp(1));
-  } else {
-    __ load_signed_short(rdx, at_bcp(1));
-  }
-  __ bswapl(rdx);
-
-  if (!is_wide) {
-    __ sarl(rdx, 16);
-  }
-  __ movl2ptr(rdx, rdx);
-
-  // Handle all the JSR stuff here, then exit.
-  // It's much shorter and cleaner than intermingling with the non-JSR
-  // normal-branch stuff occurring below.
-  if (is_jsr) {
-    // Pre-load the next target bytecode into rbx
-    __ load_unsigned_byte(rbx, Address(r13, rdx, Address::times_1, 0));
-
-    // compute return address as bci in rax
-    __ lea(rax, at_bcp((is_wide ? 5 : 3) -
-                        in_bytes(ConstMethod::codes_offset())));
-    __ subptr(rax, Address(rcx, Method::const_offset()));
-    // Adjust the bcp in r13 by the displacement in rdx
-    __ addptr(r13, rdx);
-    // jsr returns atos that is not an oop
-    __ push_i(rax);
-    __ dispatch_only(vtos);
-    return;
-  }
-
-  // Normal (non-jsr) branch handling
-
-  // Adjust the bcp in r13 by the displacement in rdx
-  __ addptr(r13, rdx);
-
-  assert(UseLoopCounter || !UseOnStackReplacement,
-         "on-stack-replacement requires loop counters");
-  Label backedge_counter_overflow;
-  Label profile_method;
-  Label dispatch;
-  if (UseLoopCounter) {
-    // increment backedge counter for backward branches
-    // rax: MDO
-    // ebx: MDO bumped taken-count
-    // rcx: method
-    // rdx: target offset
-    // r13: target bcp
-    // r14: locals pointer
-    __ testl(rdx, rdx);             // check if forward or backward branch
-    __ jcc(Assembler::positive, dispatch); // count only if backward branch
-
-    // check if MethodCounters exists
-    Label has_counters;
-    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
-    __ testptr(rax, rax);
-    __ jcc(Assembler::notZero, has_counters);
-    __ push(rdx);
-    __ push(rcx);
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::build_method_counters),
-               rcx);
-    __ pop(rcx);
-    __ pop(rdx);
-    __ movptr(rax, Address(rcx, Method::method_counters_offset()));
-    __ jcc(Assembler::zero, dispatch);
-    __ bind(has_counters);
-
-    if (TieredCompilation) {
-      Label no_mdo;
-      int increment = InvocationCounter::count_increment;
-      if (ProfileInterpreter) {
-        // Are we profiling?
-        __ movptr(rbx, Address(rcx, in_bytes(Method::method_data_offset())));
-        __ testptr(rbx, rbx);
-        __ jccb(Assembler::zero, no_mdo);
-        // Increment the MDO backedge counter
-        const Address mdo_backedge_counter(rbx, in_bytes(MethodData::backedge_counter_offset()) +
-                                           in_bytes(InvocationCounter::counter_offset()));
-        const Address mask(rbx, in_bytes(MethodData::backedge_mask_offset()));
-        __ increment_mask_and_jump(mdo_backedge_counter, increment, mask,
-                                   rax, false, Assembler::zero, &backedge_counter_overflow);
-        __ jmp(dispatch);
-      }
-      __ bind(no_mdo);
-      // Increment backedge counter in MethodCounters*
-      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
-         const Address mask(rcx, in_bytes(MethodCounters::backedge_mask_offset()));
-      __ increment_mask_and_jump(Address(rcx, be_offset), increment, mask,
-                                 rax, false, Assembler::zero, &backedge_counter_overflow);
-    } else { // not TieredCompilation
-      // increment counter
-      __ movptr(rcx, Address(rcx, Method::method_counters_offset()));
-      __ movl(rax, Address(rcx, be_offset));        // load backedge counter
-      __ incrementl(rax, InvocationCounter::count_increment); // increment counter
-      __ movl(Address(rcx, be_offset), rax);        // store counter
-
-      __ movl(rax, Address(rcx, inv_offset));    // load invocation counter
-
-      __ andl(rax, InvocationCounter::count_mask_value); // and the status bits
-      __ addl(rax, Address(rcx, be_offset));        // add both counters
-
-      if (ProfileInterpreter) {
-        // Test to see if we should create a method data oop
-        __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_profile_limit_offset())));
-        __ jcc(Assembler::less, dispatch);
-
-        // if no method data exists, go to profile method
-        __ test_method_data_pointer(rax, profile_method);
-
-        if (UseOnStackReplacement) {
-          // check for overflow against ebx which is the MDO taken count
-          __ cmp32(rbx, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
-          __ jcc(Assembler::below, dispatch);
-
-          // When ProfileInterpreter is on, the backedge_count comes
-          // from the MethodData*, which value does not get reset on
-          // the call to frequency_counter_overflow().  To avoid
-          // excessive calls to the overflow routine while the method is
-          // being compiled, add a second test to make sure the overflow
-          // function is called only once every overflow_frequency.
-          const int overflow_frequency = 1024;
-          __ andl(rbx, overflow_frequency - 1);
-          __ jcc(Assembler::zero, backedge_counter_overflow);
-
-        }
-      } else {
-        if (UseOnStackReplacement) {
-          // check for overflow against eax, which is the sum of the
-          // counters
-          __ cmp32(rax, Address(rcx, in_bytes(MethodCounters::interpreter_backward_branch_limit_offset())));
-          __ jcc(Assembler::aboveEqual, backedge_counter_overflow);
-
-        }
-      }
-    }
-    __ bind(dispatch);
-  }
-
-  // Pre-load the next target bytecode into rbx
-  __ load_unsigned_byte(rbx, Address(r13, 0));
-
-  // continue with the bytecode @ target
-  // eax: return bci for jsr's, unused otherwise
-  // ebx: target bytecode
-  // r13: target bcp
-  __ dispatch_only(vtos);
-
-  if (UseLoopCounter) {
-    if (ProfileInterpreter) {
-      // Out-of-line code to allocate method data oop.
-      __ bind(profile_method);
-      __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::profile_method));
-      __ load_unsigned_byte(rbx, Address(r13, 0));  // restore target bytecode
-      __ set_method_data_pointer_for_bcp();
-      __ jmp(dispatch);
-    }
-
-    if (UseOnStackReplacement) {
-      // invocation counter overflow
-      __ bind(backedge_counter_overflow);
-      __ negptr(rdx);
-      __ addptr(rdx, r13); // branch bcp
-      // IcoResult frequency_counter_overflow([JavaThread*], address branch_bcp)
-      __ call_VM(noreg,
-                 CAST_FROM_FN_PTR(address,
-                                  InterpreterRuntime::frequency_counter_overflow),
-                 rdx);
-      __ load_unsigned_byte(rbx, Address(r13, 0));  // restore target bytecode
-
-      // rax: osr nmethod (osr ok) or NULL (osr not possible)
-      // ebx: target bytecode
-      // rdx: scratch
-      // r14: locals pointer
-      // r13: bcp
-      __ testptr(rax, rax);                        // test result
-      __ jcc(Assembler::zero, dispatch);         // no osr if null
-      // nmethod may have been invalidated (VM may block upon call_VM return)
-      __ cmpb(Address(rax, nmethod::state_offset()), nmethod::in_use);
-      __ jcc(Assembler::notEqual, dispatch);
-
-      // We have the address of an on stack replacement routine in eax
-      // We need to prepare to execute the OSR method. First we must
-      // migrate the locals and monitors off of the stack.
-
-      __ mov(r13, rax);                             // save the nmethod
-
-      call_VM(noreg, CAST_FROM_FN_PTR(address, SharedRuntime::OSR_migration_begin));
-
-      // eax is OSR buffer, move it to expected parameter location
-      __ mov(j_rarg0, rax);
-
-      // We use j_rarg definitions here so that registers don't conflict as parameter
-      // registers change across platforms as we are in the midst of a calling
-      // sequence to the OSR nmethod and we don't want collision. These are NOT parameters.
-
-      const Register retaddr = j_rarg2;
-      const Register sender_sp = j_rarg1;
-
-      // pop the interpreter frame
-      __ movptr(sender_sp, Address(rbp, frame::interpreter_frame_sender_sp_offset * wordSize)); // get sender sp
-      __ leave();                                // remove frame anchor
-      __ pop(retaddr);                           // get return address
-      __ mov(rsp, sender_sp);                   // set sp to sender sp
-      // Ensure compiled code always sees stack at proper alignment
-      __ andptr(rsp, -(StackAlignmentInBytes));
-
-      // unlike x86 we need no specialized return from compiled code
-      // to the interpreter or the call stub.
-
-      // push the return address
-      __ push(retaddr);
-
-      // and begin the OSR nmethod
-      __ jmp(Address(r13, nmethod::osr_entry_point_offset()));
-    }
-  }
-}
-
-
-void TemplateTable::if_0cmp(Condition cc) {
-  transition(itos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ testl(rax, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-void TemplateTable::if_icmp(Condition cc) {
-  transition(itos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ pop_i(rdx);
-  __ cmpl(rdx, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-void TemplateTable::if_nullcmp(Condition cc) {
-  transition(atos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ testptr(rax, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-void TemplateTable::if_acmp(Condition cc) {
-  transition(atos, vtos);
-  // assume branch is more often taken than not (loops use backward branches)
-  Label not_taken;
-  __ pop_ptr(rdx);
-  __ cmpptr(rdx, rax);
-  __ jcc(j_not(cc), not_taken);
-  branch(false, false);
-  __ bind(not_taken);
-  __ profile_not_taken_branch(rax);
-}
-
-void TemplateTable::ret() {
-  transition(vtos, vtos);
-  locals_index(rbx);
-  __ movslq(rbx, iaddress(rbx)); // get return bci, compute return bcp
-  __ profile_ret(rbx, rcx);
-  __ get_method(rax);
-  __ movptr(r13, Address(rax, Method::const_offset()));
-  __ lea(r13, Address(r13, rbx, Address::times_1,
-                      ConstMethod::codes_offset()));
-  __ dispatch_next(vtos);
-}
-
-void TemplateTable::wide_ret() {
-  transition(vtos, vtos);
-  locals_index_wide(rbx);
-  __ movptr(rbx, aaddress(rbx)); // get return bci, compute return bcp
-  __ profile_ret(rbx, rcx);
-  __ get_method(rax);
-  __ movptr(r13, Address(rax, Method::const_offset()));
-  __ lea(r13, Address(r13, rbx, Address::times_1, ConstMethod::codes_offset()));
-  __ dispatch_next(vtos);
-}
-
-void TemplateTable::tableswitch() {
-  Label default_case, continue_execution;
-  transition(itos, vtos);
-  // align r13
-  __ lea(rbx, at_bcp(BytesPerInt));
-  __ andptr(rbx, -BytesPerInt);
-  // load lo & hi
-  __ movl(rcx, Address(rbx, BytesPerInt));
-  __ movl(rdx, Address(rbx, 2 * BytesPerInt));
-  __ bswapl(rcx);
-  __ bswapl(rdx);
-  // check against lo & hi
-  __ cmpl(rax, rcx);
-  __ jcc(Assembler::less, default_case);
-  __ cmpl(rax, rdx);
-  __ jcc(Assembler::greater, default_case);
-  // lookup dispatch offset
-  __ subl(rax, rcx);
-  __ movl(rdx, Address(rbx, rax, Address::times_4, 3 * BytesPerInt));
-  __ profile_switch_case(rax, rbx, rcx);
-  // continue execution
-  __ bind(continue_execution);
-  __ bswapl(rdx);
-  __ movl2ptr(rdx, rdx);
-  __ load_unsigned_byte(rbx, Address(r13, rdx, Address::times_1));
-  __ addptr(r13, rdx);
-  __ dispatch_only(vtos);
-  // handle default
-  __ bind(default_case);
-  __ profile_switch_default(rax);
-  __ movl(rdx, Address(rbx, 0));
-  __ jmp(continue_execution);
-}
-
-void TemplateTable::lookupswitch() {
-  transition(itos, itos);
-  __ stop("lookupswitch bytecode should have been rewritten");
-}
-
-void TemplateTable::fast_linearswitch() {
-  transition(itos, vtos);
-  Label loop_entry, loop, found, continue_execution;
-  // bswap rax so we can avoid bswapping the table entries
-  __ bswapl(rax);
-  // align r13
-  __ lea(rbx, at_bcp(BytesPerInt)); // btw: should be able to get rid of
-                                    // this instruction (change offsets
-                                    // below)
-  __ andptr(rbx, -BytesPerInt);
-  // set counter
-  __ movl(rcx, Address(rbx, BytesPerInt));
-  __ bswapl(rcx);
-  __ jmpb(loop_entry);
-  // table search
-  __ bind(loop);
-  __ cmpl(rax, Address(rbx, rcx, Address::times_8, 2 * BytesPerInt));
-  __ jcc(Assembler::equal, found);
-  __ bind(loop_entry);
-  __ decrementl(rcx);
-  __ jcc(Assembler::greaterEqual, loop);
-  // default case
-  __ profile_switch_default(rax);
-  __ movl(rdx, Address(rbx, 0));
-  __ jmp(continue_execution);
-  // entry found -> get offset
-  __ bind(found);
-  __ movl(rdx, Address(rbx, rcx, Address::times_8, 3 * BytesPerInt));
-  __ profile_switch_case(rcx, rax, rbx);
-  // continue execution
-  __ bind(continue_execution);
-  __ bswapl(rdx);
-  __ movl2ptr(rdx, rdx);
-  __ load_unsigned_byte(rbx, Address(r13, rdx, Address::times_1));
-  __ addptr(r13, rdx);
-  __ dispatch_only(vtos);
-}
-
-void TemplateTable::fast_binaryswitch() {
-  transition(itos, vtos);
-  // Implementation using the following core algorithm:
-  //
-  // int binary_search(int key, LookupswitchPair* array, int n) {
-  //   // Binary search according to "Methodik des Programmierens" by
-  //   // Edsger W. Dijkstra and W.H.J. Feijen, Addison Wesley Germany 1985.
-  //   int i = 0;
-  //   int j = n;
-  //   while (i+1 < j) {
-  //     // invariant P: 0 <= i < j <= n and (a[i] <= key < a[j] or Q)
-  //     // with      Q: for all i: 0 <= i < n: key < a[i]
-  //     // where a stands for the array and assuming that the (inexisting)
-  //     // element a[n] is infinitely big.
-  //     int h = (i + j) >> 1;
-  //     // i < h < j
-  //     if (key < array[h].fast_match()) {
-  //       j = h;
-  //     } else {
-  //       i = h;
-  //     }
-  //   }
-  //   // R: a[i] <= key < a[i+1] or Q
-  //   // (i.e., if key is within array, i is the correct index)
-  //   return i;
-  // }
-
-  // Register allocation
-  const Register key   = rax; // already set (tosca)
-  const Register array = rbx;
-  const Register i     = rcx;
-  const Register j     = rdx;
-  const Register h     = rdi;
-  const Register temp  = rsi;
-
-  // Find array start
-  __ lea(array, at_bcp(3 * BytesPerInt)); // btw: should be able to
-                                          // get rid of this
-                                          // instruction (change
-                                          // offsets below)
-  __ andptr(array, -BytesPerInt);
-
-  // Initialize i & j
-  __ xorl(i, i);                            // i = 0;
-  __ movl(j, Address(array, -BytesPerInt)); // j = length(array);
-
-  // Convert j into native byteordering
-  __ bswapl(j);
-
-  // And start
-  Label entry;
-  __ jmp(entry);
-
-  // binary search loop
-  {
-    Label loop;
-    __ bind(loop);
-    // int h = (i + j) >> 1;
-    __ leal(h, Address(i, j, Address::times_1)); // h = i + j;
-    __ sarl(h, 1);                               // h = (i + j) >> 1;
-    // if (key < array[h].fast_match()) {
-    //   j = h;
-    // } else {
-    //   i = h;
-    // }
-    // Convert array[h].match to native byte-ordering before compare
-    __ movl(temp, Address(array, h, Address::times_8));
-    __ bswapl(temp);
-    __ cmpl(key, temp);
-    // j = h if (key <  array[h].fast_match())
-    __ cmovl(Assembler::less, j, h);
-    // i = h if (key >= array[h].fast_match())
-    __ cmovl(Assembler::greaterEqual, i, h);
-    // while (i+1 < j)
-    __ bind(entry);
-    __ leal(h, Address(i, 1)); // i+1
-    __ cmpl(h, j);             // i+1 < j
-    __ jcc(Assembler::less, loop);
-  }
-
-  // end of binary search, result index is i (must check again!)
-  Label default_case;
-  // Convert array[i].match to native byte-ordering before compare
-  __ movl(temp, Address(array, i, Address::times_8));
-  __ bswapl(temp);
-  __ cmpl(key, temp);
-  __ jcc(Assembler::notEqual, default_case);
-
-  // entry found -> j = offset
-  __ movl(j , Address(array, i, Address::times_8, BytesPerInt));
-  __ profile_switch_case(i, key, array);
-  __ bswapl(j);
-  __ movl2ptr(j, j);
-  __ load_unsigned_byte(rbx, Address(r13, j, Address::times_1));
-  __ addptr(r13, j);
-  __ dispatch_only(vtos);
-
-  // default case -> j = default offset
-  __ bind(default_case);
-  __ profile_switch_default(i);
-  __ movl(j, Address(array, -2 * BytesPerInt));
-  __ bswapl(j);
-  __ movl2ptr(j, j);
-  __ load_unsigned_byte(rbx, Address(r13, j, Address::times_1));
-  __ addptr(r13, j);
-  __ dispatch_only(vtos);
-}
-
-
-void TemplateTable::_return(TosState state) {
-  transition(state, state);
-  assert(_desc->calls_vm(),
-         "inconsistent calls_vm information"); // call in remove_activation
-
-  if (_desc->bytecode() == Bytecodes::_return_register_finalizer) {
-    assert(state == vtos, "only valid state");
-    __ movptr(c_rarg1, aaddress(0));
-    __ load_klass(rdi, c_rarg1);
-    __ movl(rdi, Address(rdi, Klass::access_flags_offset()));
-    __ testl(rdi, JVM_ACC_HAS_FINALIZER);
-    Label skip_register_finalizer;
-    __ jcc(Assembler::zero, skip_register_finalizer);
-
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::register_finalizer), c_rarg1);
-
-    __ bind(skip_register_finalizer);
-  }
-
-  __ remove_activation(state, r13);
-  __ jmp(r13);
-}
-
-// ----------------------------------------------------------------------------
-// Volatile variables demand their effects be made known to all CPU's
-// in order.  Store buffers on most chips allow reads & writes to
-// reorder; the JMM's ReadAfterWrite.java test fails in -Xint mode
-// without some kind of memory barrier (i.e., it's not sufficient that
-// the interpreter does not reorder volatile references, the hardware
-// also must not reorder them).
-//
-// According to the new Java Memory Model (JMM):
-// (1) All volatiles are serialized wrt to each other.  ALSO reads &
-//     writes act as aquire & release, so:
-// (2) A read cannot let unrelated NON-volatile memory refs that
-//     happen after the read float up to before the read.  It's OK for
-//     non-volatile memory refs that happen before the volatile read to
-//     float down below it.
-// (3) Similar a volatile write cannot let unrelated NON-volatile
-//     memory refs that happen BEFORE the write float down to after the
-//     write.  It's OK for non-volatile memory refs that happen after the
-//     volatile write to float up before it.
-//
-// We only put in barriers around volatile refs (they are expensive),
-// not _between_ memory refs (that would require us to track the
-// flavor of the previous memory refs).  Requirements (2) and (3)
-// require some barriers before volatile stores and after volatile
-// loads.  These nearly cover requirement (1) but miss the
-// volatile-store-volatile-load case.  This final case is placed after
-// volatile-stores although it could just as well go before
-// volatile-loads.
-void TemplateTable::volatile_barrier(Assembler::Membar_mask_bits
-                                     order_constraint) {
-  // Helper function to insert a is-volatile test and memory barrier
-  if (os::is_MP()) { // Not needed on single CPU
-    __ membar(order_constraint);
-  }
-}
-
-void TemplateTable::resolve_cache_and_index(int byte_no,
-                                            Register Rcache,
-                                            Register index,
-                                            size_t index_size) {
-  const Register temp = rbx;
-  assert_different_registers(Rcache, index, temp);
-
-  Label resolved;
-    assert(byte_no == f1_byte || byte_no == f2_byte, "byte_no out of range");
-    __ get_cache_and_index_and_bytecode_at_bcp(Rcache, index, temp, byte_no, 1, index_size);
-    __ cmpl(temp, (int) bytecode());  // have we resolved this bytecode?
-    __ jcc(Assembler::equal, resolved);
-
-  // resolve first time through
-  address entry;
-  switch (bytecode()) {
-  case Bytecodes::_getstatic:
-  case Bytecodes::_putstatic:
-  case Bytecodes::_getfield:
-  case Bytecodes::_putfield:
-    entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_get_put);
-    break;
-  case Bytecodes::_invokevirtual:
-  case Bytecodes::_invokespecial:
-  case Bytecodes::_invokestatic:
-  case Bytecodes::_invokeinterface:
-    entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invoke);
-    break;
-  case Bytecodes::_invokehandle:
-    entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokehandle);
-    break;
-  case Bytecodes::_invokedynamic:
-    entry = CAST_FROM_FN_PTR(address, InterpreterRuntime::resolve_invokedynamic);
-    break;
-  default:
-    fatal(err_msg("unexpected bytecode: %s", Bytecodes::name(bytecode())));
-    break;
-  }
-  __ movl(temp, (int) bytecode());
-  __ call_VM(noreg, entry, temp);
-
-  // Update registers with resolved info
-  __ get_cache_and_index_at_bcp(Rcache, index, 1, index_size);
-  __ bind(resolved);
-}
-
-// The cache and index registers must be set before call
-void TemplateTable::load_field_cp_cache_entry(Register obj,
-                                              Register cache,
-                                              Register index,
-                                              Register off,
-                                              Register flags,
-                                              bool is_static = false) {
-  assert_different_registers(cache, index, flags, off);
-
-  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
-  // Field offset
-  __ movptr(off, Address(cache, index, Address::times_ptr,
-                         in_bytes(cp_base_offset +
-                                  ConstantPoolCacheEntry::f2_offset())));
-  // Flags
-  __ movl(flags, Address(cache, index, Address::times_ptr,
-                         in_bytes(cp_base_offset +
-                                  ConstantPoolCacheEntry::flags_offset())));
-
-  // klass overwrite register
-  if (is_static) {
-    __ movptr(obj, Address(cache, index, Address::times_ptr,
-                           in_bytes(cp_base_offset +
-                                    ConstantPoolCacheEntry::f1_offset())));
-    const int mirror_offset = in_bytes(Klass::java_mirror_offset());
-    __ movptr(obj, Address(obj, mirror_offset));
-  }
-}
-
-void TemplateTable::load_invoke_cp_cache_entry(int byte_no,
-                                               Register method,
-                                               Register itable_index,
-                                               Register flags,
-                                               bool is_invokevirtual,
-                                               bool is_invokevfinal, /*unused*/
-                                               bool is_invokedynamic) {
-  // setup registers
-  const Register cache = rcx;
-  const Register index = rdx;
-  assert_different_registers(method, flags);
-  assert_different_registers(method, cache, index);
-  assert_different_registers(itable_index, flags);
-  assert_different_registers(itable_index, cache, index);
-  // determine constant pool cache field offsets
-  assert(is_invokevirtual == (byte_no == f2_byte), "is_invokevirtual flag redundant");
-  const int method_offset = in_bytes(
-    ConstantPoolCache::base_offset() +
-      ((byte_no == f2_byte)
-       ? ConstantPoolCacheEntry::f2_offset()
-       : ConstantPoolCacheEntry::f1_offset()));
-  const int flags_offset = in_bytes(ConstantPoolCache::base_offset() +
-                                    ConstantPoolCacheEntry::flags_offset());
-  // access constant pool cache fields
-  const int index_offset = in_bytes(ConstantPoolCache::base_offset() +
-                                    ConstantPoolCacheEntry::f2_offset());
-
-  size_t index_size = (is_invokedynamic ? sizeof(u4) : sizeof(u2));
-  resolve_cache_and_index(byte_no, cache, index, index_size);
-    __ movptr(method, Address(cache, index, Address::times_ptr, method_offset));
-
-  if (itable_index != noreg) {
-    // pick up itable or appendix index from f2 also:
-    __ movptr(itable_index, Address(cache, index, Address::times_ptr, index_offset));
-  }
-  __ movl(flags, Address(cache, index, Address::times_ptr, flags_offset));
-}
-
-// Correct values of the cache and index registers are preserved.
-void TemplateTable::jvmti_post_field_access(Register cache, Register index,
-                                            bool is_static, bool has_tos) {
-  // do the JVMTI work here to avoid disturbing the register state below
-  // We use c_rarg registers here because we want to use the register used in
-  // the call to the VM
-  if (JvmtiExport::can_post_field_access()) {
-    // Check to see if a field access watch has been set before we
-    // take the time to call into the VM.
-    Label L1;
-    assert_different_registers(cache, index, rax);
-    __ mov32(rax, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
-    __ testl(rax, rax);
-    __ jcc(Assembler::zero, L1);
-
-    __ get_cache_and_index_at_bcp(c_rarg2, c_rarg3, 1);
-
-    // cache entry pointer
-    __ addptr(c_rarg2, in_bytes(ConstantPoolCache::base_offset()));
-    __ shll(c_rarg3, LogBytesPerWord);
-    __ addptr(c_rarg2, c_rarg3);
-    if (is_static) {
-      __ xorl(c_rarg1, c_rarg1); // NULL object reference
-    } else {
-      __ movptr(c_rarg1, at_tos()); // get object pointer without popping it
-      __ verify_oop(c_rarg1);
-    }
-    // c_rarg1: object pointer or NULL
-    // c_rarg2: cache entry pointer
-    // c_rarg3: jvalue object on the stack
-    __ call_VM(noreg, CAST_FROM_FN_PTR(address,
-                                       InterpreterRuntime::post_field_access),
-               c_rarg1, c_rarg2, c_rarg3);
-    __ get_cache_and_index_at_bcp(cache, index, 1);
-    __ bind(L1);
-  }
-}
-
-void TemplateTable::pop_and_check_object(Register r) {
-  __ pop_ptr(r);
-  __ null_check(r);  // for field access must check obj.
-  __ verify_oop(r);
-}
-
-void TemplateTable::getfield_or_static(int byte_no, bool is_static) {
-  transition(vtos, vtos);
-
-  const Register cache = rcx;
-  const Register index = rdx;
-  const Register obj   = c_rarg3;
-  const Register off   = rbx;
-  const Register flags = rax;
-  const Register bc = c_rarg3; // uses same reg as obj, so don't mix them
-
-  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
-  jvmti_post_field_access(cache, index, is_static, false);
-  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
-
-  if (!is_static) {
-    // obj is on the stack
-    pop_and_check_object(obj);
-  }
-
-  const Address field(obj, off, Address::times_1);
-
-  Label Done, notByte, notInt, notShort, notChar,
-              notLong, notFloat, notObj, notDouble;
-
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-  // Make sure we don't need to mask edx after the above shift
-  assert(btos == 0, "change code, btos != 0");
-
-  __ andl(flags, ConstantPoolCacheEntry::tos_state_mask);
-  __ jcc(Assembler::notZero, notByte);
-  // btos
-  __ load_signed_byte(rax, field);
-  __ push(btos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_bgetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notByte);
-  __ cmpl(flags, atos);
-  __ jcc(Assembler::notEqual, notObj);
-  // atos
-  __ load_heap_oop(rax, field);
-  __ push(atos);
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_agetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notObj);
-  __ cmpl(flags, itos);
-  __ jcc(Assembler::notEqual, notInt);
-  // itos
-  __ movl(rax, field);
-  __ push(itos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_igetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notInt);
-  __ cmpl(flags, ctos);
-  __ jcc(Assembler::notEqual, notChar);
-  // ctos
-  __ load_unsigned_short(rax, field);
-  __ push(ctos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_cgetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notChar);
-  __ cmpl(flags, stos);
-  __ jcc(Assembler::notEqual, notShort);
-  // stos
-  __ load_signed_short(rax, field);
-  __ push(stos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_sgetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notShort);
-  __ cmpl(flags, ltos);
-  __ jcc(Assembler::notEqual, notLong);
-  // ltos
-  __ movq(rax, field);
-  __ push(ltos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_lgetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notLong);
-  __ cmpl(flags, ftos);
-  __ jcc(Assembler::notEqual, notFloat);
-  // ftos
-  __ movflt(xmm0, field);
-  __ push(ftos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_fgetfield, bc, rbx);
-  }
-  __ jmp(Done);
-
-  __ bind(notFloat);
-#ifdef ASSERT
-  __ cmpl(flags, dtos);
-  __ jcc(Assembler::notEqual, notDouble);
-#endif
-  // dtos
-  __ movdbl(xmm0, field);
-  __ push(dtos);
-  // Rewrite bytecode to be faster
-  if (!is_static) {
-    patch_bytecode(Bytecodes::_fast_dgetfield, bc, rbx);
-  }
-#ifdef ASSERT
-  __ jmp(Done);
-
-  __ bind(notDouble);
-  __ stop("Bad state");
-#endif
-
-  __ bind(Done);
-  // [jk] not needed currently
-  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadLoad |
-  //                                              Assembler::LoadStore));
-}
-
-
-void TemplateTable::getfield(int byte_no) {
-  getfield_or_static(byte_no, false);
-}
-
-void TemplateTable::getstatic(int byte_no) {
-  getfield_or_static(byte_no, true);
-}
-
-// The registers cache and index expected to be set before call.
-// The function may destroy various registers, just not the cache and index registers.
-void TemplateTable::jvmti_post_field_mod(Register cache, Register index, bool is_static) {
-  transition(vtos, vtos);
-
-  ByteSize cp_base_offset = ConstantPoolCache::base_offset();
-
-  if (JvmtiExport::can_post_field_modification()) {
-    // Check to see if a field modification watch has been set before
-    // we take the time to call into the VM.
-    Label L1;
-    assert_different_registers(cache, index, rax);
-    __ mov32(rax, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
-    __ testl(rax, rax);
-    __ jcc(Assembler::zero, L1);
-
-    __ get_cache_and_index_at_bcp(c_rarg2, rscratch1, 1);
-
-    if (is_static) {
-      // Life is simple.  Null out the object pointer.
-      __ xorl(c_rarg1, c_rarg1);
-    } else {
-      // Life is harder. The stack holds the value on top, followed by
-      // the object.  We don't know the size of the value, though; it
-      // could be one or two words depending on its type. As a result,
-      // we must find the type to determine where the object is.
-      __ movl(c_rarg3, Address(c_rarg2, rscratch1,
-                           Address::times_8,
-                           in_bytes(cp_base_offset +
-                                     ConstantPoolCacheEntry::flags_offset())));
-      __ shrl(c_rarg3, ConstantPoolCacheEntry::tos_state_shift);
-      // Make sure we don't need to mask rcx after the above shift
-      ConstantPoolCacheEntry::verify_tos_state_shift();
-      __ movptr(c_rarg1, at_tos_p1());  // initially assume a one word jvalue
-      __ cmpl(c_rarg3, ltos);
-      __ cmovptr(Assembler::equal,
-                 c_rarg1, at_tos_p2()); // ltos (two word jvalue)
-      __ cmpl(c_rarg3, dtos);
-      __ cmovptr(Assembler::equal,
-                 c_rarg1, at_tos_p2()); // dtos (two word jvalue)
-    }
-    // cache entry pointer
-    __ addptr(c_rarg2, in_bytes(cp_base_offset));
-    __ shll(rscratch1, LogBytesPerWord);
-    __ addptr(c_rarg2, rscratch1);
-    // object (tos)
-    __ mov(c_rarg3, rsp);
-    // c_rarg1: object pointer set up above (NULL if static)
-    // c_rarg2: cache entry pointer
-    // c_rarg3: jvalue object on the stack
-    __ call_VM(noreg,
-               CAST_FROM_FN_PTR(address,
-                                InterpreterRuntime::post_field_modification),
-               c_rarg1, c_rarg2, c_rarg3);
-    __ get_cache_and_index_at_bcp(cache, index, 1);
-    __ bind(L1);
-  }
-}
-
-void TemplateTable::putfield_or_static(int byte_no, bool is_static) {
-  transition(vtos, vtos);
-
-  const Register cache = rcx;
-  const Register index = rdx;
-  const Register obj   = rcx;
-  const Register off   = rbx;
-  const Register flags = rax;
-  const Register bc    = c_rarg3;
-
-  resolve_cache_and_index(byte_no, cache, index, sizeof(u2));
-  jvmti_post_field_mod(cache, index, is_static);
-  load_field_cp_cache_entry(obj, cache, index, off, flags, is_static);
-
-  // [jk] not needed currently
-  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadStore |
-  //                                              Assembler::StoreStore));
-
-  Label notVolatile, Done;
-  __ movl(rdx, flags);
-  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  __ andl(rdx, 0x1);
-
-  // field address
-  const Address field(obj, off, Address::times_1);
-
-  Label notByte, notInt, notShort, notChar,
-        notLong, notFloat, notObj, notDouble;
-
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-
-  assert(btos == 0, "change code, btos != 0");
-  __ andl(flags, ConstantPoolCacheEntry::tos_state_mask);
-  __ jcc(Assembler::notZero, notByte);
-
-  // btos
-  {
-    __ pop(btos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movb(field, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_bputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notByte);
-  __ cmpl(flags, atos);
-  __ jcc(Assembler::notEqual, notObj);
-
-  // atos
-  {
-    __ pop(atos);
-    if (!is_static) pop_and_check_object(obj);
-    // Store into the field
-    do_oop_store(_masm, field, rax, _bs->kind(), false);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_aputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notObj);
-  __ cmpl(flags, itos);
-  __ jcc(Assembler::notEqual, notInt);
-
-  // itos
-  {
-    __ pop(itos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movl(field, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_iputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notInt);
-  __ cmpl(flags, ctos);
-  __ jcc(Assembler::notEqual, notChar);
-
-  // ctos
-  {
-    __ pop(ctos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movw(field, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_cputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notChar);
-  __ cmpl(flags, stos);
-  __ jcc(Assembler::notEqual, notShort);
-
-  // stos
-  {
-    __ pop(stos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movw(field, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_sputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notShort);
-  __ cmpl(flags, ltos);
-  __ jcc(Assembler::notEqual, notLong);
-
-  // ltos
-  {
-    __ pop(ltos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movq(field, rax);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_lputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notLong);
-  __ cmpl(flags, ftos);
-  __ jcc(Assembler::notEqual, notFloat);
-
-  // ftos
-  {
-    __ pop(ftos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movflt(field, xmm0);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_fputfield, bc, rbx, true, byte_no);
-    }
-    __ jmp(Done);
-  }
-
-  __ bind(notFloat);
-#ifdef ASSERT
-  __ cmpl(flags, dtos);
-  __ jcc(Assembler::notEqual, notDouble);
-#endif
-
-  // dtos
-  {
-    __ pop(dtos);
-    if (!is_static) pop_and_check_object(obj);
-    __ movdbl(field, xmm0);
-    if (!is_static) {
-      patch_bytecode(Bytecodes::_fast_dputfield, bc, rbx, true, byte_no);
-    }
-  }
-
-#ifdef ASSERT
-  __ jmp(Done);
-
-  __ bind(notDouble);
-  __ stop("Bad state");
-#endif
-
-  __ bind(Done);
-
-  // Check for volatile store
-  __ testl(rdx, rdx);
-  __ jcc(Assembler::zero, notVolatile);
-  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
-                                               Assembler::StoreStore));
-  __ bind(notVolatile);
-}
-
-void TemplateTable::putfield(int byte_no) {
-  putfield_or_static(byte_no, false);
-}
-
-void TemplateTable::putstatic(int byte_no) {
-  putfield_or_static(byte_no, true);
-}
-
-void TemplateTable::jvmti_post_fast_field_mod() {
-  if (JvmtiExport::can_post_field_modification()) {
-    // Check to see if a field modification watch has been set before
-    // we take the time to call into the VM.
-    Label L2;
-    __ mov32(c_rarg3, ExternalAddress((address)JvmtiExport::get_field_modification_count_addr()));
-    __ testl(c_rarg3, c_rarg3);
-    __ jcc(Assembler::zero, L2);
-    __ pop_ptr(rbx);                  // copy the object pointer from tos
-    __ verify_oop(rbx);
-    __ push_ptr(rbx);                 // put the object pointer back on tos
-    // Save tos values before call_VM() clobbers them. Since we have
-    // to do it for every data type, we use the saved values as the
-    // jvalue object.
-    switch (bytecode()) {          // load values into the jvalue object
-    case Bytecodes::_fast_aputfield: __ push_ptr(rax); break;
-    case Bytecodes::_fast_bputfield: // fall through
-    case Bytecodes::_fast_sputfield: // fall through
-    case Bytecodes::_fast_cputfield: // fall through
-    case Bytecodes::_fast_iputfield: __ push_i(rax); break;
-    case Bytecodes::_fast_dputfield: __ push_d(); break;
-    case Bytecodes::_fast_fputfield: __ push_f(); break;
-    case Bytecodes::_fast_lputfield: __ push_l(rax); break;
-
-    default:
-      ShouldNotReachHere();
-    }
-    __ mov(c_rarg3, rsp);             // points to jvalue on the stack
-    // access constant pool cache entry
-    __ get_cache_entry_pointer_at_bcp(c_rarg2, rax, 1);
-    __ verify_oop(rbx);
-    // rbx: object pointer copied above
-    // c_rarg2: cache entry pointer
-    // c_rarg3: jvalue object on the stack
-    __ call_VM(noreg,
-               CAST_FROM_FN_PTR(address,
-                                InterpreterRuntime::post_field_modification),
-               rbx, c_rarg2, c_rarg3);
-
-    switch (bytecode()) {             // restore tos values
-    case Bytecodes::_fast_aputfield: __ pop_ptr(rax); break;
-    case Bytecodes::_fast_bputfield: // fall through
-    case Bytecodes::_fast_sputfield: // fall through
-    case Bytecodes::_fast_cputfield: // fall through
-    case Bytecodes::_fast_iputfield: __ pop_i(rax); break;
-    case Bytecodes::_fast_dputfield: __ pop_d(); break;
-    case Bytecodes::_fast_fputfield: __ pop_f(); break;
-    case Bytecodes::_fast_lputfield: __ pop_l(rax); break;
-    }
-    __ bind(L2);
-  }
-}
-
-void TemplateTable::fast_storefield(TosState state) {
-  transition(state, vtos);
-
-  ByteSize base = ConstantPoolCache::base_offset();
-
-  jvmti_post_fast_field_mod();
-
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
-
-  // test for volatile with rdx
-  __ movl(rdx, Address(rcx, rbx, Address::times_8,
-                       in_bytes(base +
-                                ConstantPoolCacheEntry::flags_offset())));
-
-  // replace index with field offset from cache entry
-  __ movptr(rbx, Address(rcx, rbx, Address::times_8,
-                         in_bytes(base + ConstantPoolCacheEntry::f2_offset())));
-
-  // [jk] not needed currently
-  // volatile_barrier(Assembler::Membar_mask_bits(Assembler::LoadStore |
-  //                                              Assembler::StoreStore));
-
-  Label notVolatile;
-  __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  __ andl(rdx, 0x1);
-
-  // Get object from stack
-  pop_and_check_object(rcx);
-
-  // field address
-  const Address field(rcx, rbx, Address::times_1);
-
-  // access field
-  switch (bytecode()) {
-  case Bytecodes::_fast_aputfield:
-    do_oop_store(_masm, field, rax, _bs->kind(), false);
-    break;
-  case Bytecodes::_fast_lputfield:
-    __ movq(field, rax);
-    break;
-  case Bytecodes::_fast_iputfield:
-    __ movl(field, rax);
-    break;
-  case Bytecodes::_fast_bputfield:
-    __ movb(field, rax);
-    break;
-  case Bytecodes::_fast_sputfield:
-    // fall through
-  case Bytecodes::_fast_cputfield:
-    __ movw(field, rax);
-    break;
-  case Bytecodes::_fast_fputfield:
-    __ movflt(field, xmm0);
-    break;
-  case Bytecodes::_fast_dputfield:
-    __ movdbl(field, xmm0);
-    break;
-  default:
-    ShouldNotReachHere();
-  }
-
-  // Check for volatile store
-  __ testl(rdx, rdx);
-  __ jcc(Assembler::zero, notVolatile);
-  volatile_barrier(Assembler::Membar_mask_bits(Assembler::StoreLoad |
-                                               Assembler::StoreStore));
-  __ bind(notVolatile);
-}
-
-
-void TemplateTable::fast_accessfield(TosState state) {
-  transition(atos, state);
-
-  // Do the JVMTI work here to avoid disturbing the register state below
-  if (JvmtiExport::can_post_field_access()) {
-    // Check to see if a field access watch has been set before we
-    // take the time to call into the VM.
-    Label L1;
-    __ mov32(rcx, ExternalAddress((address) JvmtiExport::get_field_access_count_addr()));
-    __ testl(rcx, rcx);
-    __ jcc(Assembler::zero, L1);
-    // access constant pool cache entry
-    __ get_cache_entry_pointer_at_bcp(c_rarg2, rcx, 1);
-    __ verify_oop(rax);
-    __ push_ptr(rax);  // save object pointer before call_VM() clobbers it
-    __ mov(c_rarg1, rax);
-    // c_rarg1: object pointer copied above
-    // c_rarg2: cache entry pointer
-    __ call_VM(noreg,
-               CAST_FROM_FN_PTR(address,
-                                InterpreterRuntime::post_field_access),
-               c_rarg1, c_rarg2);
-    __ pop_ptr(rax); // restore object pointer
-    __ bind(L1);
-  }
-
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rbx, 1);
-  // replace index with field offset from cache entry
-  // [jk] not needed currently
-  // if (os::is_MP()) {
-  //   __ movl(rdx, Address(rcx, rbx, Address::times_8,
-  //                        in_bytes(ConstantPoolCache::base_offset() +
-  //                                 ConstantPoolCacheEntry::flags_offset())));
-  //   __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  //   __ andl(rdx, 0x1);
-  // }
-  __ movptr(rbx, Address(rcx, rbx, Address::times_8,
-                         in_bytes(ConstantPoolCache::base_offset() +
-                                  ConstantPoolCacheEntry::f2_offset())));
-
-  // rax: object
-  __ verify_oop(rax);
-  __ null_check(rax);
-  Address field(rax, rbx, Address::times_1);
-
-  // access field
-  switch (bytecode()) {
-  case Bytecodes::_fast_agetfield:
-    __ load_heap_oop(rax, field);
-    __ verify_oop(rax);
-    break;
-  case Bytecodes::_fast_lgetfield:
-    __ movq(rax, field);
-    break;
-  case Bytecodes::_fast_igetfield:
-    __ movl(rax, field);
-    break;
-  case Bytecodes::_fast_bgetfield:
-    __ movsbl(rax, field);
-    break;
-  case Bytecodes::_fast_sgetfield:
-    __ load_signed_short(rax, field);
-    break;
-  case Bytecodes::_fast_cgetfield:
-    __ load_unsigned_short(rax, field);
-    break;
-  case Bytecodes::_fast_fgetfield:
-    __ movflt(xmm0, field);
-    break;
-  case Bytecodes::_fast_dgetfield:
-    __ movdbl(xmm0, field);
-    break;
-  default:
-    ShouldNotReachHere();
-  }
-  // [jk] not needed currently
-  // if (os::is_MP()) {
-  //   Label notVolatile;
-  //   __ testl(rdx, rdx);
-  //   __ jcc(Assembler::zero, notVolatile);
-  //   __ membar(Assembler::LoadLoad);
-  //   __ bind(notVolatile);
-  //};
-}
-
-void TemplateTable::fast_xaccess(TosState state) {
-  transition(vtos, state);
-
-  // get receiver
-  __ movptr(rax, aaddress(0));
-  // access constant pool cache
-  __ get_cache_and_index_at_bcp(rcx, rdx, 2);
-  __ movptr(rbx,
-            Address(rcx, rdx, Address::times_8,
-                    in_bytes(ConstantPoolCache::base_offset() +
-                             ConstantPoolCacheEntry::f2_offset())));
-  // make sure exception is reported in correct bcp range (getfield is
-  // next instruction)
-  __ increment(r13);
-  __ null_check(rax);
-  switch (state) {
-  case itos:
-    __ movl(rax, Address(rax, rbx, Address::times_1));
-    break;
-  case atos:
-    __ load_heap_oop(rax, Address(rax, rbx, Address::times_1));
-    __ verify_oop(rax);
-    break;
-  case ftos:
-    __ movflt(xmm0, Address(rax, rbx, Address::times_1));
-    break;
-  default:
-    ShouldNotReachHere();
-  }
-
-  // [jk] not needed currently
-  // if (os::is_MP()) {
-  //   Label notVolatile;
-  //   __ movl(rdx, Address(rcx, rdx, Address::times_8,
-  //                        in_bytes(ConstantPoolCache::base_offset() +
-  //                                 ConstantPoolCacheEntry::flags_offset())));
-  //   __ shrl(rdx, ConstantPoolCacheEntry::is_volatile_shift);
-  //   __ testl(rdx, 0x1);
-  //   __ jcc(Assembler::zero, notVolatile);
-  //   __ membar(Assembler::LoadLoad);
-  //   __ bind(notVolatile);
-  // }
-
-  __ decrement(r13);
-}
-
-
-
-//-----------------------------------------------------------------------------
-// Calls
-
-void TemplateTable::count_calls(Register method, Register temp) {
-  // implemented elsewhere
-  ShouldNotReachHere();
-}
-
-void TemplateTable::prepare_invoke(int byte_no,
-                                   Register method,  // linked method (or i-klass)
-                                   Register index,   // itable index, MethodType, etc.
-                                   Register recv,    // if caller wants to see it
-                                   Register flags    // if caller wants to test it
-                                   ) {
-  // determine flags
-  const Bytecodes::Code code = bytecode();
-  const bool is_invokeinterface  = code == Bytecodes::_invokeinterface;
-  const bool is_invokedynamic    = code == Bytecodes::_invokedynamic;
-  const bool is_invokehandle     = code == Bytecodes::_invokehandle;
-  const bool is_invokevirtual    = code == Bytecodes::_invokevirtual;
-  const bool is_invokespecial    = code == Bytecodes::_invokespecial;
-  const bool load_receiver       = (recv  != noreg);
-  const bool save_flags          = (flags != noreg);
-  assert(load_receiver == (code != Bytecodes::_invokestatic && code != Bytecodes::_invokedynamic), "");
-  assert(save_flags    == (is_invokeinterface || is_invokevirtual), "need flags for vfinal");
-  assert(flags == noreg || flags == rdx, "");
-  assert(recv  == noreg || recv  == rcx, "");
-
-  // setup registers & access constant pool cache
-  if (recv  == noreg)  recv  = rcx;
-  if (flags == noreg)  flags = rdx;
-  assert_different_registers(method, index, recv, flags);
-
-  // save 'interpreter return address'
-  __ save_bcp();
-
-  load_invoke_cp_cache_entry(byte_no, method, index, flags, is_invokevirtual, false, is_invokedynamic);
-
-  // maybe push appendix to arguments (just before return address)
-  if (is_invokedynamic || is_invokehandle) {
-    Label L_no_push;
-    __ testl(flags, (1 << ConstantPoolCacheEntry::has_appendix_shift));
-    __ jcc(Assembler::zero, L_no_push);
-    // Push the appendix as a trailing parameter.
-    // This must be done before we get the receiver,
-    // since the parameter_size includes it.
-    __ push(rbx);
-    __ mov(rbx, index);
-    assert(ConstantPoolCacheEntry::_indy_resolved_references_appendix_offset == 0, "appendix expected at index+0");
-    __ load_resolved_reference_at_index(index, rbx);
-    __ pop(rbx);
-    __ push(index);  // push appendix (MethodType, CallSite, etc.)
-    __ bind(L_no_push);
-  }
-
-  // load receiver if needed (after appendix is pushed so parameter size is correct)
-  // Note: no return address pushed yet
-  if (load_receiver) {
-    __ movl(recv, flags);
-    __ andl(recv, ConstantPoolCacheEntry::parameter_size_mask);
-    const int no_return_pc_pushed_yet = -1;  // argument slot correction before we push return address
-    const int receiver_is_at_end      = -1;  // back off one slot to get receiver
-    Address recv_addr = __ argument_address(recv, no_return_pc_pushed_yet + receiver_is_at_end);
-    __ movptr(recv, recv_addr);
-    __ verify_oop(recv);
-  }
-
-  if (save_flags) {
-    __ movl(r13, flags);
-  }
-
-  // compute return type
-  __ shrl(flags, ConstantPoolCacheEntry::tos_state_shift);
-  // Make sure we don't need to mask flags after the above shift
-  ConstantPoolCacheEntry::verify_tos_state_shift();
-  // load return address
-  {
-    const address table_addr = (address) Interpreter::invoke_return_entry_table_for(code);
-    ExternalAddress table(table_addr);
-    __ lea(rscratch1, table);
-    __ movptr(flags, Address(rscratch1, flags, Address::times_ptr));
-  }
-
-  // push return address
-  __ push(flags);
-
-  // Restore flags value from the constant pool cache, and restore rsi
-  // for later null checks.  r13 is the bytecode pointer
-  if (save_flags) {
-    __ movl(flags, r13);
-    __ restore_bcp();
-  }
-}
-
-
-void TemplateTable::invokevirtual_helper(Register index,
-                                         Register recv,
-                                         Register flags) {
-  // Uses temporary registers rax, rdx
-  assert_different_registers(index, recv, rax, rdx);
-  assert(index == rbx, "");
-  assert(recv  == rcx, "");
-
-  // Test for an invoke of a final method
-  Label notFinal;
-  __ movl(rax, flags);
-  __ andl(rax, (1 << ConstantPoolCacheEntry::is_vfinal_shift));
-  __ jcc(Assembler::zero, notFinal);
-
-  const Register method = index;  // method must be rbx
-  assert(method == rbx,
-         "Method* must be rbx for interpreter calling convention");
-
-  // do the call - the index is actually the method to call
-  // that is, f2 is a vtable index if !is_vfinal, else f2 is a Method*
-
-  // It's final, need a null check here!
-  __ null_check(recv);
-
-  // profile this call
-  __ profile_final_call(rax);
-  __ profile_arguments_type(rax, method, r13, true);
-
-  __ jump_from_interpreted(method, rax);
-
-  __ bind(notFinal);
-
-  // get receiver klass
-  __ null_check(recv, oopDesc::klass_offset_in_bytes());
-  __ load_klass(rax, recv);
-
-  // profile this call
-  __ profile_virtual_call(rax, r14, rdx);
-
-  // get target Method* & entry point
-  __ lookup_virtual_method(rax, index, method);
-  __ profile_arguments_type(rdx, method, r13, true);
-  __ jump_from_interpreted(method, rdx);
-}
-
-
-void TemplateTable::invokevirtual(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f2_byte, "use this argument");
-  prepare_invoke(byte_no,
-                 rbx,    // method or vtable index
-                 noreg,  // unused itable index
-                 rcx, rdx); // recv, flags
-
-  // rbx: index
-  // rcx: receiver
-  // rdx: flags
-
-  invokevirtual_helper(rbx, rcx, rdx);
-}
-
-
-void TemplateTable::invokespecial(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rbx, noreg,  // get f1 Method*
-                 rcx);  // get receiver also for null check
-  __ verify_oop(rcx);
-  __ null_check(rcx);
-  // do the call
-  __ profile_call(rax);
-  __ profile_arguments_type(rax, rbx, r13, false);
-  __ jump_from_interpreted(rbx, rax);
-}
-
-
-void TemplateTable::invokestatic(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rbx);  // get f1 Method*
-  // do the call
-  __ profile_call(rax);
-  __ profile_arguments_type(rax, rbx, r13, false);
-  __ jump_from_interpreted(rbx, rax);
-}
-
-void TemplateTable::fast_invokevfinal(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f2_byte, "use this argument");
-  __ stop("fast_invokevfinal not used on amd64");
-}
-
-void TemplateTable::invokeinterface(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  prepare_invoke(byte_no, rax, rbx,  // get f1 Klass*, f2 itable index
-                 rcx, rdx); // recv, flags
-
-  // rax: interface klass (from f1)
-  // rbx: itable index (from f2)
-  // rcx: receiver
-  // rdx: flags
-
-  // Special case of invokeinterface called for virtual method of
-  // java.lang.Object.  See cpCacheOop.cpp for details.
-  // This code isn't produced by javac, but could be produced by
-  // another compliant java compiler.
-  Label notMethod;
-  __ movl(r14, rdx);
-  __ andl(r14, (1 << ConstantPoolCacheEntry::is_forced_virtual_shift));
-  __ jcc(Assembler::zero, notMethod);
-
-  invokevirtual_helper(rbx, rcx, rdx);
-  __ bind(notMethod);
-
-  // Get receiver klass into rdx - also a null check
-  __ restore_locals();  // restore r14
-  __ null_check(rcx, oopDesc::klass_offset_in_bytes());
-  __ load_klass(rdx, rcx);
-
-  // profile this call
-  __ profile_virtual_call(rdx, r13, r14);
-
-  Label no_such_interface, no_such_method;
-
-  __ lookup_interface_method(// inputs: rec. class, interface, itable index
-                             rdx, rax, rbx,
-                             // outputs: method, scan temp. reg
-                             rbx, r13,
-                             no_such_interface);
-
-  // rbx: Method* to call
-  // rcx: receiver
-  // Check for abstract method error
-  // Note: This should be done more efficiently via a throw_abstract_method_error
-  //       interpreter entry point and a conditional jump to it in case of a null
-  //       method.
-  __ testptr(rbx, rbx);
-  __ jcc(Assembler::zero, no_such_method);
-
-  __ profile_arguments_type(rdx, rbx, r13, true);
-
-  // do the call
-  // rcx: receiver
-  // rbx,: Method*
-  __ jump_from_interpreted(rbx, rdx);
-  __ should_not_reach_here();
-
-  // exception handling code follows...
-  // note: must restore interpreter registers to canonical
-  //       state for exception handling to work correctly!
-
-  __ bind(no_such_method);
-  // throw exception
-  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
-  __ restore_bcp();      // r13 must be correct for exception handler   (was destroyed)
-  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::throw_AbstractMethodError));
-  // the call_VM checks for exception, so we should never return here.
-  __ should_not_reach_here();
-
-  __ bind(no_such_interface);
-  // throw exception
-  __ pop(rbx);           // pop return address (pushed by prepare_invoke)
-  __ restore_bcp();      // r13 must be correct for exception handler   (was destroyed)
-  __ restore_locals();   // make sure locals pointer is correct as well (was destroyed)
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address,
-                   InterpreterRuntime::throw_IncompatibleClassChangeError));
-  // the call_VM checks for exception, so we should never return here.
-  __ should_not_reach_here();
-}
-
-
-void TemplateTable::invokehandle(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-  const Register rbx_method = rbx;
-  const Register rax_mtype  = rax;
-  const Register rcx_recv   = rcx;
-  const Register rdx_flags  = rdx;
-
-  prepare_invoke(byte_no, rbx_method, rax_mtype, rcx_recv);
-  __ verify_method_ptr(rbx_method);
-  __ verify_oop(rcx_recv);
-  __ null_check(rcx_recv);
-
-  // rax: MethodType object (from cpool->resolved_references[f1], if necessary)
-  // rbx: MH.invokeExact_MT method (from f2)
-
-  // Note:  rax_mtype is already pushed (if necessary) by prepare_invoke
-
-  // FIXME: profile the LambdaForm also
-  __ profile_final_call(rax);
-  __ profile_arguments_type(rdx, rbx_method, r13, true);
-
-  __ jump_from_interpreted(rbx_method, rdx);
-}
-
-
-void TemplateTable::invokedynamic(int byte_no) {
-  transition(vtos, vtos);
-  assert(byte_no == f1_byte, "use this argument");
-
-  const Register rbx_method   = rbx;
-  const Register rax_callsite = rax;
-
-  prepare_invoke(byte_no, rbx_method, rax_callsite);
-
-  // rax: CallSite object (from cpool->resolved_references[f1])
-  // rbx: MH.linkToCallSite method (from f2)
-
-  // Note:  rax_callsite is already pushed by prepare_invoke
-
-  // %%% should make a type profile for any invokedynamic that takes a ref argument
-  // profile this call
-  __ profile_call(r13);
-  __ profile_arguments_type(rdx, rbx_method, r13, false);
-
-  __ verify_oop(rax_callsite);
-
-  __ jump_from_interpreted(rbx_method, rdx);
-}
-
-
-//-----------------------------------------------------------------------------
-// Allocation
-
-void TemplateTable::_new() {
-  transition(vtos, atos);
-  __ get_unsigned_2_byte_index_at_bcp(rdx, 1);
-  Label slow_case;
-  Label done;
-  Label initialize_header;
-  Label initialize_object; // including clearing the fields
-  Label allocate_shared;
-
-  __ get_cpool_and_tags(rsi, rax);
-  // Make sure the class we're about to instantiate has been resolved.
-  // This is done before loading InstanceKlass to be consistent with the order
-  // how Constant Pool is updated (see ConstantPool::klass_at_put)
-  const int tags_offset = Array<u1>::base_offset_in_bytes();
-  __ cmpb(Address(rax, rdx, Address::times_1, tags_offset),
-          JVM_CONSTANT_Class);
-  __ jcc(Assembler::notEqual, slow_case);
-
-  // get InstanceKlass
-  __ movptr(rsi, Address(rsi, rdx,
-            Address::times_8, sizeof(ConstantPool)));
-
-  // make sure klass is initialized & doesn't have finalizer
-  // make sure klass is fully initialized
-  __ cmpb(Address(rsi,
-                  InstanceKlass::init_state_offset()),
-          InstanceKlass::fully_initialized);
-  __ jcc(Assembler::notEqual, slow_case);
-
-  // get instance_size in InstanceKlass (scaled to a count of bytes)
-  __ movl(rdx,
-          Address(rsi,
-                  Klass::layout_helper_offset()));
-  // test to see if it has a finalizer or is malformed in some way
-  __ testl(rdx, Klass::_lh_instance_slow_path_bit);
-  __ jcc(Assembler::notZero, slow_case);
-
-  // Allocate the instance
-  // 1) Try to allocate in the TLAB
-  // 2) if fail and the object is large allocate in the shared Eden
-  // 3) if the above fails (or is not applicable), go to a slow case
-  // (creates a new TLAB, etc.)
-
-  const bool allow_shared_alloc =
-    Universe::heap()->supports_inline_contig_alloc();
-
-  if (UseTLAB) {
-    __ movptr(rax, Address(r15_thread, in_bytes(JavaThread::tlab_top_offset())));
-    __ lea(rbx, Address(rax, rdx, Address::times_1));
-    __ cmpptr(rbx, Address(r15_thread, in_bytes(JavaThread::tlab_end_offset())));
-    __ jcc(Assembler::above, allow_shared_alloc ? allocate_shared : slow_case);
-    __ movptr(Address(r15_thread, in_bytes(JavaThread::tlab_top_offset())), rbx);
-    if (ZeroTLAB) {
-      // the fields have been already cleared
-      __ jmp(initialize_header);
-    } else {
-      // initialize both the header and fields
-      __ jmp(initialize_object);
-    }
-  }
-
-  // Allocation in the shared Eden, if allowed.
-  //
-  // rdx: instance size in bytes
-  if (allow_shared_alloc) {
-    __ bind(allocate_shared);
-
-    ExternalAddress top((address)Universe::heap()->top_addr());
-    ExternalAddress end((address)Universe::heap()->end_addr());
-
-    const Register RtopAddr = rscratch1;
-    const Register RendAddr = rscratch2;
-
-    __ lea(RtopAddr, top);
-    __ lea(RendAddr, end);
-    __ movptr(rax, Address(RtopAddr, 0));
-
-    // For retries rax gets set by cmpxchgq
-    Label retry;
-    __ bind(retry);
-    __ lea(rbx, Address(rax, rdx, Address::times_1));
-    __ cmpptr(rbx, Address(RendAddr, 0));
-    __ jcc(Assembler::above, slow_case);
-
-    // Compare rax with the top addr, and if still equal, store the new
-    // top addr in rbx at the address of the top addr pointer. Sets ZF if was
-    // equal, and clears it otherwise. Use lock prefix for atomicity on MPs.
-    //
-    // rax: object begin
-    // rbx: object end
-    // rdx: instance size in bytes
-    if (os::is_MP()) {
-      __ lock();
-    }
-    __ cmpxchgptr(rbx, Address(RtopAddr, 0));
-
-    // if someone beat us on the allocation, try again, otherwise continue
-    __ jcc(Assembler::notEqual, retry);
-
-    __ incr_allocated_bytes(r15_thread, rdx, 0);
-  }
-
-  if (UseTLAB || Universe::heap()->supports_inline_contig_alloc()) {
-    // The object is initialized before the header.  If the object size is
-    // zero, go directly to the header initialization.
-    __ bind(initialize_object);
-    __ decrementl(rdx, sizeof(oopDesc));
-    __ jcc(Assembler::zero, initialize_header);
-
-    // Initialize object fields
-    __ xorl(rcx, rcx); // use zero reg to clear memory (shorter code)
-    __ shrl(rdx, LogBytesPerLong);  // divide by oopSize to simplify the loop
-    {
-      Label loop;
-      __ bind(loop);
-      __ movq(Address(rax, rdx, Address::times_8,
-                      sizeof(oopDesc) - oopSize),
-              rcx);
-      __ decrementl(rdx);
-      __ jcc(Assembler::notZero, loop);
-    }
-
-    // initialize object header only.
-    __ bind(initialize_header);
-    if (UseBiasedLocking) {
-      __ movptr(rscratch1, Address(rsi, Klass::prototype_header_offset()));
-      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes()), rscratch1);
-    } else {
-      __ movptr(Address(rax, oopDesc::mark_offset_in_bytes()),
-               (intptr_t) markOopDesc::prototype()); // header (address 0x1)
-    }
-    __ xorl(rcx, rcx); // use zero reg to clear memory (shorter code)
-    __ store_klass_gap(rax, rcx);  // zero klass gap for compressed oops
-    __ store_klass(rax, rsi);      // store klass last
-
-    {
-      SkipIfEqual skip(_masm, &DTraceAllocProbes, false);
-      // Trigger dtrace event for fastpath
-      __ push(atos); // save the return value
-      __ call_VM_leaf(
-           CAST_FROM_FN_PTR(address, SharedRuntime::dtrace_object_alloc), rax);
-      __ pop(atos); // restore the return value
-
-    }
-    __ jmp(done);
-  }
-
-
-  // slow case
-  __ bind(slow_case);
-  __ get_constant_pool(c_rarg1);
-  __ get_unsigned_2_byte_index_at_bcp(c_rarg2, 1);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::_new), c_rarg1, c_rarg2);
-  __ verify_oop(rax);
-
-  // continue
-  __ bind(done);
-}
-
-void TemplateTable::newarray() {
-  transition(itos, atos);
-  __ load_unsigned_byte(c_rarg1, at_bcp(1));
-  __ movl(c_rarg2, rax);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::newarray),
-          c_rarg1, c_rarg2);
-}
-
-void TemplateTable::anewarray() {
-  transition(itos, atos);
-  __ get_unsigned_2_byte_index_at_bcp(c_rarg2, 1);
-  __ get_constant_pool(c_rarg1);
-  __ movl(c_rarg3, rax);
-  call_VM(rax, CAST_FROM_FN_PTR(address, InterpreterRuntime::anewarray),
-          c_rarg1, c_rarg2, c_rarg3);
-}
-
-void TemplateTable::arraylength() {
-  transition(atos, itos);
-  __ null_check(rax, arrayOopDesc::length_offset_in_bytes());
-  __ movl(rax, Address(rax, arrayOopDesc::length_offset_in_bytes()));
-}
-
-void TemplateTable::checkcast() {
-  transition(atos, atos);
-  Label done, is_null, ok_is_subtype, quicked, resolved;
-  __ testptr(rax, rax); // object is in rax
-  __ jcc(Assembler::zero, is_null);
-
-  // Get cpool & tags index
-  __ get_cpool_and_tags(rcx, rdx); // rcx=cpool, rdx=tags array
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // rbx=index
-  // See if bytecode has already been quicked
-  __ cmpb(Address(rdx, rbx,
-                  Address::times_1,
-                  Array<u1>::base_offset_in_bytes()),
-          JVM_CONSTANT_Class);
-  __ jcc(Assembler::equal, quicked);
-  __ push(atos); // save receiver for result, and for GC
-  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc));
-  // vm_result_2 has metadata result
-  __ get_vm_result_2(rax, r15_thread);
-  __ pop_ptr(rdx); // restore receiver
-  __ jmpb(resolved);
-
-  // Get superklass in rax and subklass in rbx
-  __ bind(quicked);
-  __ mov(rdx, rax); // Save object in rdx; rax needed for subtype check
-  __ movptr(rax, Address(rcx, rbx,
-                       Address::times_8, sizeof(ConstantPool)));
-
-  __ bind(resolved);
-  __ load_klass(rbx, rdx);
-
-  // Generate subtype check.  Blows rcx, rdi.  Object in rdx.
-  // Superklass in rax.  Subklass in rbx.
-  __ gen_subtype_check(rbx, ok_is_subtype);
-
-  // Come here on failure
-  __ push_ptr(rdx);
-  // object is at TOS
-  __ jump(ExternalAddress(Interpreter::_throw_ClassCastException_entry));
-
-  // Come here on success
-  __ bind(ok_is_subtype);
-  __ mov(rax, rdx); // Restore object in rdx
-
-  // Collect counts on whether this check-cast sees NULLs a lot or not.
-  if (ProfileInterpreter) {
-    __ jmp(done);
-    __ bind(is_null);
-    __ profile_null_seen(rcx);
-  } else {
-    __ bind(is_null);   // same as 'done'
-  }
-  __ bind(done);
-}
-
-void TemplateTable::instanceof() {
-  transition(atos, itos);
-  Label done, is_null, ok_is_subtype, quicked, resolved;
-  __ testptr(rax, rax);
-  __ jcc(Assembler::zero, is_null);
-
-  // Get cpool & tags index
-  __ get_cpool_and_tags(rcx, rdx); // rcx=cpool, rdx=tags array
-  __ get_unsigned_2_byte_index_at_bcp(rbx, 1); // rbx=index
-  // See if bytecode has already been quicked
-  __ cmpb(Address(rdx, rbx,
-                  Address::times_1,
-                  Array<u1>::base_offset_in_bytes()),
-          JVM_CONSTANT_Class);
-  __ jcc(Assembler::equal, quicked);
-
-  __ push(atos); // save receiver for result, and for GC
-  call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::quicken_io_cc));
-  // vm_result_2 has metadata result
-  __ get_vm_result_2(rax, r15_thread);
-  __ pop_ptr(rdx); // restore receiver
-  __ verify_oop(rdx);
-  __ load_klass(rdx, rdx);
-  __ jmpb(resolved);
-
-  // Get superklass in rax and subklass in rdx
-  __ bind(quicked);
-  __ load_klass(rdx, rax);
-  __ movptr(rax, Address(rcx, rbx,
-                         Address::times_8, sizeof(ConstantPool)));
-
-  __ bind(resolved);
-
-  // Generate subtype check.  Blows rcx, rdi
-  // Superklass in rax.  Subklass in rdx.
-  __ gen_subtype_check(rdx, ok_is_subtype);
-
-  // Come here on failure
-  __ xorl(rax, rax);
-  __ jmpb(done);
-  // Come here on success
-  __ bind(ok_is_subtype);
-  __ movl(rax, 1);
-
-  // Collect counts on whether this test sees NULLs a lot or not.
-  if (ProfileInterpreter) {
-    __ jmp(done);
-    __ bind(is_null);
-    __ profile_null_seen(rcx);
-  } else {
-    __ bind(is_null);   // same as 'done'
-  }
-  __ bind(done);
-  // rax = 0: obj == NULL or  obj is not an instanceof the specified klass
-  // rax = 1: obj != NULL and obj is     an instanceof the specified klass
-}
-
-//-----------------------------------------------------------------------------
-// Breakpoints
-void TemplateTable::_breakpoint() {
-  // Note: We get here even if we are single stepping..
-  // jbug inists on setting breakpoints at every bytecode
-  // even if we are in single step mode.
-
-  transition(vtos, vtos);
-
-  // get the unpatched byte code
-  __ get_method(c_rarg1);
-  __ call_VM(noreg,
-             CAST_FROM_FN_PTR(address,
-                              InterpreterRuntime::get_original_bytecode_at),
-             c_rarg1, r13);
-  __ mov(rbx, rax);
-
-  // post the breakpoint event
-  __ get_method(c_rarg1);
-  __ call_VM(noreg,
-             CAST_FROM_FN_PTR(address, InterpreterRuntime::_breakpoint),
-             c_rarg1, r13);
-
-  // complete the execution of original bytecode
-  __ dispatch_only_normal(vtos);
-}
-
-//-----------------------------------------------------------------------------
-// Exceptions
-
-void TemplateTable::athrow() {
-  transition(atos, vtos);
-  __ null_check(rax);
-  __ jump(ExternalAddress(Interpreter::throw_exception_entry()));
-}
-
-//-----------------------------------------------------------------------------
-// Synchronization
-//
-// Note: monitorenter & exit are symmetric routines; which is reflected
-//       in the assembly code structure as well
-//
-// Stack layout:
-//
-// [expressions  ] <--- rsp               = expression stack top
-// ..
-// [expressions  ]
-// [monitor entry] <--- monitor block top = expression stack bot
-// ..
-// [monitor entry]
-// [frame data   ] <--- monitor block bot
-// ...
-// [saved rbp    ] <--- rbp
-void TemplateTable::monitorenter() {
-  transition(atos, vtos);
-
-  // check for NULL object
-  __ null_check(rax);
-
-  const Address monitor_block_top(
-        rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
-  const Address monitor_block_bot(
-        rbp, frame::interpreter_frame_initial_sp_offset * wordSize);
-  const int entry_size = frame::interpreter_frame_monitor_size() * wordSize;
-
-  Label allocated;
-
-  // initialize entry pointer
-  __ xorl(c_rarg1, c_rarg1); // points to free slot or NULL
-
-  // find a free slot in the monitor block (result in c_rarg1)
-  {
-    Label entry, loop, exit;
-    __ movptr(c_rarg3, monitor_block_top); // points to current entry,
-                                     // starting with top-most entry
-    __ lea(c_rarg2, monitor_block_bot); // points to word before bottom
-                                     // of monitor block
-    __ jmpb(entry);
-
-    __ bind(loop);
-    // check if current entry is used
-    __ cmpptr(Address(c_rarg3, BasicObjectLock::obj_offset_in_bytes()), (int32_t) NULL_WORD);
-    // if not used then remember entry in c_rarg1
-    __ cmov(Assembler::equal, c_rarg1, c_rarg3);
-    // check if current entry is for same object
-    __ cmpptr(rax, Address(c_rarg3, BasicObjectLock::obj_offset_in_bytes()));
-    // if same object then stop searching
-    __ jccb(Assembler::equal, exit);
-    // otherwise advance to next entry
-    __ addptr(c_rarg3, entry_size);
-    __ bind(entry);
-    // check if bottom reached
-    __ cmpptr(c_rarg3, c_rarg2);
-    // if not at bottom then check this entry
-    __ jcc(Assembler::notEqual, loop);
-    __ bind(exit);
-  }
-
-  __ testptr(c_rarg1, c_rarg1); // check if a slot has been found
-  __ jcc(Assembler::notZero, allocated); // if found, continue with that one
-
-  // allocate one if there's no free slot
-  {
-    Label entry, loop;
-    // 1. compute new pointers             // rsp: old expression stack top
-    __ movptr(c_rarg1, monitor_block_bot); // c_rarg1: old expression stack bottom
-    __ subptr(rsp, entry_size);            // move expression stack top
-    __ subptr(c_rarg1, entry_size);        // move expression stack bottom
-    __ mov(c_rarg3, rsp);                  // set start value for copy loop
-    __ movptr(monitor_block_bot, c_rarg1); // set new monitor block bottom
-    __ jmp(entry);
-    // 2. move expression stack contents
-    __ bind(loop);
-    __ movptr(c_rarg2, Address(c_rarg3, entry_size)); // load expression stack
-                                                      // word from old location
-    __ movptr(Address(c_rarg3, 0), c_rarg2);          // and store it at new location
-    __ addptr(c_rarg3, wordSize);                     // advance to next word
-    __ bind(entry);
-    __ cmpptr(c_rarg3, c_rarg1);            // check if bottom reached
-    __ jcc(Assembler::notEqual, loop);      // if not at bottom then
-                                            // copy next word
-  }
-
-  // call run-time routine
-  // c_rarg1: points to monitor entry
-  __ bind(allocated);
-
-  // Increment bcp to point to the next bytecode, so exception
-  // handling for async. exceptions work correctly.
-  // The object has already been poped from the stack, so the
-  // expression stack looks correct.
-  __ increment(r13);
-
-  // store object
-  __ movptr(Address(c_rarg1, BasicObjectLock::obj_offset_in_bytes()), rax);
-  __ lock_object(c_rarg1);
-
-  // check to make sure this monitor doesn't cause stack overflow after locking
-  __ save_bcp();  // in case of exception
-  __ generate_stack_overflow_check(0);
-
-  // The bcp has already been incremented. Just need to dispatch to
-  // next instruction.
-  __ dispatch_next(vtos);
-}
-
-
-void TemplateTable::monitorexit() {
-  transition(atos, vtos);
-
-  // check for NULL object
-  __ null_check(rax);
-
-  const Address monitor_block_top(
-        rbp, frame::interpreter_frame_monitor_block_top_offset * wordSize);
-  const Address monitor_block_bot(
-        rbp, frame::interpreter_frame_initial_sp_offset * wordSize);
-  const int entry_size = frame::interpreter_frame_monitor_size() * wordSize;
-
-  Label found;
-
-  // find matching slot
-  {
-    Label entry, loop;
-    __ movptr(c_rarg1, monitor_block_top); // points to current entry,
-                                     // starting with top-most entry
-    __ lea(c_rarg2, monitor_block_bot); // points to word before bottom
-                                     // of monitor block
-    __ jmpb(entry);
-
-    __ bind(loop);
-    // check if current entry is for same object
-    __ cmpptr(rax, Address(c_rarg1, BasicObjectLock::obj_offset_in_bytes()));
-    // if same object then stop searching
-    __ jcc(Assembler::equal, found);
-    // otherwise advance to next entry
-    __ addptr(c_rarg1, entry_size);
-    __ bind(entry);
-    // check if bottom reached
-    __ cmpptr(c_rarg1, c_rarg2);
-    // if not at bottom then check this entry
-    __ jcc(Assembler::notEqual, loop);
-  }
-
-  // error handling. Unlocking was not block-structured
-  __ call_VM(noreg, CAST_FROM_FN_PTR(address,
-                   InterpreterRuntime::throw_illegal_monitor_state_exception));
-  __ should_not_reach_here();
-
-  // call run-time routine
-  // rsi: points to monitor entry
-  __ bind(found);
-  __ push_ptr(rax); // make sure object is on stack (contract with oopMaps)
-  __ unlock_object(c_rarg1);
-  __ pop_ptr(rax); // discard object
-}
-
-
-// Wide instructions
-void TemplateTable::wide() {
-  transition(vtos, vtos);
-  __ load_unsigned_byte(rbx, at_bcp(1));
-  __ lea(rscratch1, ExternalAddress((address)Interpreter::_wentry_point));
-  __ jmp(Address(rscratch1, rbx, Address::times_8));
-  // Note: the r13 increment step is part of the individual wide
-  // bytecode implementations
-}
-
-
-// Multi arrays
-void TemplateTable::multianewarray() {
-  transition(vtos, atos);
-  __ load_unsigned_byte(rax, at_bcp(3)); // get number of dimensions
-  // last dim is on top of stack; we want address of first one:
-  // first_addr = last_addr + (ndims - 1) * wordSize
-  __ lea(c_rarg1, Address(rsp, rax, Address::times_8, -wordSize));
-  call_VM(rax,
-          CAST_FROM_FN_PTR(address, InterpreterRuntime::multianewarray),
-          c_rarg1);
-  __ load_unsigned_byte(rbx, at_bcp(3));
-  __ lea(rsp, Address(rsp, rbx, Address::times_8));
-}
-#endif // !CC_INTERP
--- a/hotspot/src/os/solaris/vm/jvm_solaris.h	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os/solaris/vm/jvm_solaris.h	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1998, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1998, 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
@@ -41,7 +41,9 @@
  * JNI conversion, which should be sorted out later.
  */
 
+#define __USE_LEGACY_PROTOTYPES__
 #include <dirent.h>             /* For DIR */
+#undef __USE_LEGACY_PROTOTYPES__
 #include <sys/param.h>          /* For MAXPATHLEN */
 #include <sys/socket.h>         /* For socklen_t */
 #include <unistd.h>             /* For F_OK, R_OK, W_OK */
--- a/hotspot/src/os_cpu/aix_ppc/vm/orderAccess_aix_ppc.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/aix_ppc/vm/orderAccess_aix_ppc.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2015, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2012, 2014, SAP AG. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -28,6 +28,9 @@
 
 #include "runtime/orderAccess.hpp"
 
+// Compiler version last used for testing: xlc 12
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
 //
@@ -61,86 +64,30 @@
 #define inlasm_lwsync()   __asm__ __volatile__ ("lwsync" : : : "memory");
 #define inlasm_eieio()    __asm__ __volatile__ ("eieio"  : : : "memory");
 #define inlasm_isync()    __asm__ __volatile__ ("isync"  : : : "memory");
-#define inlasm_release()  inlasm_lwsync();
-#define inlasm_acquire()  inlasm_lwsync();
 // Use twi-isync for load_acquire (faster than lwsync).
 // ATTENTION: seems like xlC 10.1 has problems with this inline assembler macro (VerifyMethodHandles found "bad vminfo in AMH.conv"):
 // #define inlasm_acquire_reg(X) __asm__ __volatile__ ("twi 0,%0,0\n isync\n" : : "r" (X) : "memory");
 #define inlasm_acquire_reg(X) inlasm_lwsync();
-#define inlasm_fence()    inlasm_sync();
 
-inline void     OrderAccess::loadload()   { inlasm_lwsync();  }
-inline void     OrderAccess::storestore() { inlasm_lwsync();  }
-inline void     OrderAccess::loadstore()  { inlasm_lwsync();  }
-inline void     OrderAccess::storeload()  { inlasm_fence();   }
-
-inline void     OrderAccess::acquire()    { inlasm_acquire(); }
-inline void     OrderAccess::release()    { inlasm_release(); }
-inline void     OrderAccess::fence()      { inlasm_fence();   }
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { register jbyte t = *p;   inlasm_acquire_reg(t); return t; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { register jshort t = *p;  inlasm_acquire_reg(t); return t; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { register jint t = *p;    inlasm_acquire_reg(t); return t; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { register jlong t = *p;   inlasm_acquire_reg(t); return t; }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { register jubyte t = *p;  inlasm_acquire_reg(t); return t; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { register jushort t = *p; inlasm_acquire_reg(t); return t; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { register juint t = *p;   inlasm_acquire_reg(t); return t; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return (julong)load_acquire((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { register jfloat t = *p;  inlasm_acquire(); return t; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { register jdouble t = *p; inlasm_acquire(); return t; }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return (intptr_t)load_acquire((volatile jlong*)p); }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return (void*)   load_acquire((volatile jlong*)p); }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return (void*)   load_acquire((volatile jlong*)p); }
+inline void OrderAccess::loadload()   { inlasm_lwsync(); }
+inline void OrderAccess::storestore() { inlasm_lwsync(); }
+inline void OrderAccess::loadstore()  { inlasm_lwsync(); }
+inline void OrderAccess::storeload()  { inlasm_sync();   }
 
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { inlasm_release(); *p = v; }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { inlasm_release(); *(void* volatile *)p = v; }
+inline void OrderAccess::acquire()    { inlasm_lwsync(); }
+inline void OrderAccess::release()    { inlasm_lwsync(); }
+inline void OrderAccess::fence()      { inlasm_sync();   }
 
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { inlasm_release(); *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { inlasm_release(); *(void* volatile *)p = v; inlasm_fence(); }
+template<> inline jbyte  OrderAccess::specialized_load_acquire<jbyte> (volatile jbyte*  p) { register jbyte t = load(p);  inlasm_acquire_reg(t); return t; }
+template<> inline jshort OrderAccess::specialized_load_acquire<jshort>(volatile jshort* p) { register jshort t = load(p); inlasm_acquire_reg(t); return t; }
+template<> inline jint   OrderAccess::specialized_load_acquire<jint>  (volatile jint*   p) { register jint t = load(p);   inlasm_acquire_reg(t); return t; }
+template<> inline jlong  OrderAccess::specialized_load_acquire<jlong> (volatile jlong*  p) { register jlong t = load(p);  inlasm_acquire_reg(t); return t; }
 
 #undef inlasm_sync
 #undef inlasm_lwsync
 #undef inlasm_eieio
 #undef inlasm_isync
-#undef inlasm_release
-#undef inlasm_acquire
-#undef inlasm_fence
+
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_AIX_OJDKPPC_VM_ORDERACCESS_AIX_PPC_INLINE_HPP
--- a/hotspot/src/os_cpu/bsd_x86/vm/orderAccess_bsd_x86.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/bsd_x86/vm/orderAccess_bsd_x86.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -29,27 +29,27 @@
 #include "runtime/orderAccess.hpp"
 #include "runtime/os.hpp"
 
+// Compiler version last used for testing: clang 5.1
+// Please update this information when this file changes
+
+// A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
+static inline void compiler_barrier() {
+  __asm__ volatile ("" : : : "memory");
+}
+
+// x86 is TSO and hence only needs a fence for storeload
+// However, a compiler barrier is still needed to prevent reordering
+// between volatile and non-volatile memory accesses.
+
 // Implementation of class OrderAccess.
 
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
+inline void OrderAccess::storeload()  { fence();            }
 
-inline void OrderAccess::acquire() {
-  volatile intptr_t local_dummy;
-#ifdef AMD64
-  __asm__ volatile ("movq 0(%%rsp), %0" : "=r" (local_dummy) : : "memory");
-#else
-  __asm__ volatile ("movl 0(%%esp),%0" : "=r" (local_dummy) : : "memory");
-#endif // AMD64
-}
-
-inline void OrderAccess::release() {
-  // Avoid hitting the same cache-line from
-  // different threads.
-  volatile jint local_dummy = 0;
-}
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
 inline void OrderAccess::fence() {
   if (os::is_MP()) {
@@ -60,156 +60,50 @@
     __asm__ volatile ("lock; addl $0,0(%%esp)" : : : "cc", "memory");
 #endif
   }
+  compiler_barrier();
 }
 
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return *p; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return *p; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { return *p; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return Atomic::load(p); }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return *p; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return *p; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { return *p; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return Atomic::load((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return *p; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return jdouble_cast(Atomic::load((volatile jlong*)p)); }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return *p; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return *(void* volatile *)p; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return *(void* const volatile *)p; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { Atomic::store(v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { Atomic::store((jlong)v, (volatile jlong*)p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { release_store((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*  p, jbyte  v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jbyte> (volatile jbyte*  p, jbyte  v) {
   __asm__ volatile (  "xchgb (%2),%0"
                     : "=q" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::store_fence(jshort* p, jshort v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jshort>(volatile jshort* p, jshort v) {
   __asm__ volatile (  "xchgw (%2),%0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::store_fence(jint*   p, jint   v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jint>  (volatile jint*   p, jint   v) {
   __asm__ volatile (  "xchgl (%2),%0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
 
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) {
 #ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  *p = v; fence();
-#endif // AMD64
-}
-
-// AMD64 copied the bodies for the the signed version. 32bit did this. As long as the
-// compiler does the inlining this is simpler.
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { store_fence((jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { store_fence((jshort*)p, (jshort)v); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { store_fence((jint*)p,   (jint)v);   }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { store_fence((jlong*)p,  (jlong)v);  }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; fence(); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
-}
-
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) {
-#ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
-}
-
-// Must duplicate definitions instead of calling store_fence because we don't want to cast away volatile.
-inline void     OrderAccess::release_store_fence(volatile jbyte*  p, jbyte  v) {
-  __asm__ volatile (  "xchgb (%2),%0"
-                    : "=q" (v)
-                    : "0" (v), "r" (p)
-                    : "memory");
-}
-inline void     OrderAccess::release_store_fence(volatile jshort* p, jshort v) {
-  __asm__ volatile (  "xchgw (%2),%0"
+template<>
+inline void OrderAccess::specialized_release_store_fence<jlong> (volatile jlong*  p, jlong  v) {
+  __asm__ volatile (  "xchgq (%2), %0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::release_store_fence(volatile jint*   p, jint   v) {
-  __asm__ volatile (  "xchgl (%2),%0"
-                    : "=r" (v)
-                    : "0" (v), "r" (p)
-                    : "memory");
-}
+#endif // AMD64
 
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store(p, v); fence();
-#endif // AMD64
+template<>
+inline void OrderAccess::specialized_release_store_fence<jfloat> (volatile jfloat*  p, jfloat  v) {
+  release_store_fence((volatile jint*)p, jint_cast(v));
+}
+template<>
+inline void OrderAccess::specialized_release_store_fence<jdouble>(volatile jdouble* p, jdouble v) {
+  release_store_fence((volatile jlong*)p, jlong_cast(v));
 }
 
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { release_store_fence((volatile jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { release_store_fence((volatile jshort*)p, (jshort)v); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { release_store_fence((volatile jint*)p,   (jint)v);   }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store_fence((volatile jlong*)p,  (jlong)v);  }
-
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store_fence((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_BSD_X86_VM_ORDERACCESS_BSD_X86_INLINE_HPP
--- a/hotspot/src/os_cpu/bsd_zero/vm/orderAccess_bsd_zero.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/bsd_zero/vm/orderAccess_bsd_zero.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 2015, Oracle and/or its affiliates. All rights reserved.
  * Copyright 2007, 2008, 2009 Red Hat, Inc.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -40,8 +40,7 @@
 #define __kernel_dmb (*(__kernel_dmb_t *) 0xffff0fa0)
 
 #define FULL_MEM_BARRIER __kernel_dmb()
-#define READ_MEM_BARRIER __kernel_dmb()
-#define WRITE_MEM_BARRIER __kernel_dmb()
+#define LIGHT_MEM_BARRIER __kernel_dmb()
 
 #else // ARM
 
@@ -50,126 +49,31 @@
 #ifdef PPC
 
 #ifdef __NO_LWSYNC__
-#define READ_MEM_BARRIER __asm __volatile ("sync":::"memory")
-#define WRITE_MEM_BARRIER __asm __volatile ("sync":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("sync":::"memory")
 #else
-#define READ_MEM_BARRIER __asm __volatile ("lwsync":::"memory")
-#define WRITE_MEM_BARRIER __asm __volatile ("lwsync":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("lwsync":::"memory")
 #endif
 
 #else // PPC
 
-#define READ_MEM_BARRIER __asm __volatile ("":::"memory")
-#define WRITE_MEM_BARRIER __asm __volatile ("":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("":::"memory")
 
 #endif // PPC
 
 #endif // ARM
 
-
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
-
-inline void OrderAccess::acquire() {
-  READ_MEM_BARRIER;
-}
-
-inline void OrderAccess::release() {
-  WRITE_MEM_BARRIER;
-}
-
-inline void OrderAccess::fence() {
-  FULL_MEM_BARRIER;
-}
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { jbyte data = *p; acquire(); return data; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { jshort data = *p; acquire(); return data; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { jint data = *p; acquire(); return data; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) {
-  jlong tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
-inline jubyte    OrderAccess::load_acquire(volatile jubyte*   p) { jubyte data = *p; acquire(); return data; }
-inline jushort   OrderAccess::load_acquire(volatile jushort*  p) { jushort data = *p; acquire(); return data; }
-inline juint     OrderAccess::load_acquire(volatile juint*    p) { juint data = *p; acquire(); return data; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) {
-  julong tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { jfloat data = *p; acquire(); return data; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) {
-  jdouble tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
+// Note: What is meant by LIGHT_MEM_BARRIER is a barrier which is sufficient
+// to provide TSO semantics, i.e. StoreStore | LoadLoad | LoadStore.
 
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) {
-  intptr_t data = *p;
-  acquire();
-  return data;
-}
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) {
-  void *data = *(void* volatile *)p;
-  acquire();
-  return data;
-}
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) {
-  void *data = *(void* const volatile *)p;
-  acquire();
-  return data;
-}
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v)
-{ release(); os::atomic_copy64(&v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v)
-{ release(); os::atomic_copy64(&v, p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v)
-{ release(); os::atomic_copy64(&v, p); }
+inline void OrderAccess::loadload()   { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::storestore() { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::loadstore()  { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::storeload()  { FULL_MEM_BARRIER;  }
 
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { release(); *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v)
-{ release(); *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { os::atomic_copy64(&v, p); fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { os::atomic_copy64(&v, p); fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { os::atomic_copy64(&v, p); fence(); }
+inline void OrderAccess::acquire()    { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::release()    { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::fence()      { FULL_MEM_BARRIER;  }
 
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; fence(); }
-
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store(p, v); fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { release_store_ptr(p, v); fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { release_store_ptr(p, v); fence(); }
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_BSD_ZERO_VM_ORDERACCESS_BSD_ZERO_INLINE_HPP
--- a/hotspot/src/os_cpu/linux_ppc/vm/orderAccess_linux_ppc.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/linux_ppc/vm/orderAccess_linux_ppc.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 2015, Oracle and/or its affiliates. All rights reserved.
  * Copyright (c) 2012, 2014, SAP AG. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -32,6 +32,9 @@
 #error "OrderAccess currently only implemented for PPC64"
 #endif
 
+// Compiler version last used for testing: gcc 4.1.2
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
 //
@@ -65,84 +68,29 @@
 #define inlasm_lwsync()   __asm__ __volatile__ ("lwsync" : : : "memory");
 #define inlasm_eieio()    __asm__ __volatile__ ("eieio"  : : : "memory");
 #define inlasm_isync()    __asm__ __volatile__ ("isync"  : : : "memory");
-#define inlasm_release()  inlasm_lwsync();
-#define inlasm_acquire()  inlasm_lwsync();
 // Use twi-isync for load_acquire (faster than lwsync).
 #define inlasm_acquire_reg(X) __asm__ __volatile__ ("twi 0,%0,0\n isync\n" : : "r" (X) : "memory");
-#define inlasm_fence()    inlasm_sync();
 
-inline void     OrderAccess::loadload()   { inlasm_lwsync();  }
-inline void     OrderAccess::storestore() { inlasm_lwsync();  }
-inline void     OrderAccess::loadstore()  { inlasm_lwsync();  }
-inline void     OrderAccess::storeload()  { inlasm_fence();   }
-
-inline void     OrderAccess::acquire()    { inlasm_acquire(); }
-inline void     OrderAccess::release()    { inlasm_release(); }
-inline void     OrderAccess::fence()      { inlasm_fence();   }
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { register jbyte t = *p;   inlasm_acquire_reg(t); return t; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { register jshort t = *p;  inlasm_acquire_reg(t); return t; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { register jint t = *p;    inlasm_acquire_reg(t); return t; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { register jlong t = *p;   inlasm_acquire_reg(t); return t; }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { register jubyte t = *p;  inlasm_acquire_reg(t); return t; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { register jushort t = *p; inlasm_acquire_reg(t); return t; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { register juint t = *p;   inlasm_acquire_reg(t); return t; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return (julong)load_acquire((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { register jfloat t = *p;  inlasm_acquire(); return t; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { register jdouble t = *p; inlasm_acquire(); return t; }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return (intptr_t)load_acquire((volatile jlong*)p); }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return (void*)   load_acquire((volatile jlong*)p); }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return (void*)   load_acquire((volatile jlong*)p); }
+inline void   OrderAccess::loadload()   { inlasm_lwsync(); }
+inline void   OrderAccess::storestore() { inlasm_lwsync(); }
+inline void   OrderAccess::loadstore()  { inlasm_lwsync(); }
+inline void   OrderAccess::storeload()  { inlasm_sync();   }
 
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { inlasm_release(); *p = v; }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { inlasm_release(); *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { inlasm_release(); *(void* volatile *)p = v; }
+inline void   OrderAccess::acquire()    { inlasm_lwsync(); }
+inline void   OrderAccess::release()    { inlasm_lwsync(); }
+inline void   OrderAccess::fence()      { inlasm_sync();   }
 
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; inlasm_fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { inlasm_release(); *p = v; inlasm_fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { inlasm_release(); *p = v; inlasm_fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { inlasm_release(); *(void* volatile *)p = v; inlasm_fence(); }
+template<> inline jbyte  OrderAccess::specialized_load_acquire<jbyte> (volatile jbyte*  p) { register jbyte t = load(p);  inlasm_acquire_reg(t); return t; }
+template<> inline jshort OrderAccess::specialized_load_acquire<jshort>(volatile jshort* p) { register jshort t = load(p); inlasm_acquire_reg(t); return t; }
+template<> inline jint   OrderAccess::specialized_load_acquire<jint>  (volatile jint*   p) { register jint t = load(p);   inlasm_acquire_reg(t); return t; }
+template<> inline jlong  OrderAccess::specialized_load_acquire<jlong> (volatile jlong*  p) { register jlong t = load(p);  inlasm_acquire_reg(t); return t; }
 
 #undef inlasm_sync
 #undef inlasm_lwsync
 #undef inlasm_eieio
 #undef inlasm_isync
-#undef inlasm_release
-#undef inlasm_acquire
-#undef inlasm_fence
+#undef inlasm_acquire_reg
+
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_LINUX_PPC_VM_ORDERACCESS_LINUX_PPC_INLINE_HPP
--- a/hotspot/src/os_cpu/linux_sparc/vm/orderAccess_linux_sparc.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/linux_sparc/vm/orderAccess_linux_sparc.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -29,81 +29,25 @@
 
 // Implementation of class OrderAccess.
 
-// Assume TSO.
-
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
-
-inline void OrderAccess::acquire() {
-  __asm__ volatile ("nop" : : :);
-}
-
-inline void OrderAccess::release() {
-  jint* local_dummy = (jint*)&local_dummy;
-  __asm__ volatile("stw %%g0, [%0]" : : "r" (local_dummy) : "memory");
-}
-
-inline void OrderAccess::fence() {
-  __asm__ volatile ("membar  #StoreLoad" : : :);
+// A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
+static inline void compiler_barrier() {
+  __asm__ volatile ("" : : : "memory");
 }
 
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return *p; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return *p; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { return *p; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return *p; }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return *p; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return *p; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { return *p; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return *p; }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return *p; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return *p; }
+// Assume TSO.
 
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return *p; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return *(void* volatile *)p; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return *(void* const volatile *)p; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { *p = v; }
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
+inline void OrderAccess::storeload()  { fence();            }
 
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; fence(); }
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; fence(); }
+inline void OrderAccess::fence() {
+  __asm__ volatile ("membar  #StoreLoad" : : : "memory");
+}
 
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { *p = v; fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { *(void* volatile *)p = v; fence(); }
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_LINUX_SPARC_VM_ORDERACCESS_LINUX_SPARC_INLINE_HPP
--- a/hotspot/src/os_cpu/linux_x86/vm/orderAccess_linux_x86.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/linux_x86/vm/orderAccess_linux_x86.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -29,6 +29,9 @@
 #include "runtime/orderAccess.hpp"
 #include "runtime/os.hpp"
 
+// Compiler version last used for testing: gcc 4.8.2
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
 // A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
@@ -36,23 +39,13 @@
   __asm__ volatile ("" : : : "memory");
 }
 
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
+inline void OrderAccess::storeload()  { fence();            }
 
-inline void OrderAccess::acquire() {
-  volatile intptr_t local_dummy;
-#ifdef AMD64
-  __asm__ volatile ("movq 0(%%rsp), %0" : "=r" (local_dummy) : : "memory");
-#else
-  __asm__ volatile ("movl 0(%%esp),%0" : "=r" (local_dummy) : : "memory");
-#endif // AMD64
-}
-
-inline void OrderAccess::release() {
-  compiler_barrier();
-}
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
 inline void OrderAccess::fence() {
   if (os::is_MP()) {
@@ -63,156 +56,50 @@
     __asm__ volatile ("lock; addl $0,0(%%esp)" : : : "cc", "memory");
 #endif
   }
+  compiler_barrier();
 }
 
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { jbyte   v = *p; compiler_barrier(); return v; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { jshort  v = *p; compiler_barrier(); return v; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { jint    v = *p; compiler_barrier(); return v; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { jlong   v = Atomic::load(p); compiler_barrier(); return v; }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { jubyte  v = *p; compiler_barrier(); return v; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { jushort v = *p; compiler_barrier(); return v; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { juint   v = *p; compiler_barrier(); return v; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { julong  v = Atomic::load((volatile jlong*)p); compiler_barrier(); return v; }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { jfloat  v = *p; compiler_barrier(); return v; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { jdouble v = jdouble_cast(Atomic::load((volatile jlong*)p)); compiler_barrier(); return v; }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { intptr_t v = *p; compiler_barrier(); return v; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { void*    v = *(void* volatile *)p; compiler_barrier(); return v; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { void*    v = *(void* const volatile *)p; compiler_barrier(); return v; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { compiler_barrier(); Atomic::store(v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { compiler_barrier(); Atomic::store((jlong)v, (volatile jlong*)p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { release_store((volatile jlong *)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { compiler_barrier(); *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { compiler_barrier(); *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*  p, jbyte  v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jbyte> (volatile jbyte*  p, jbyte  v) {
   __asm__ volatile (  "xchgb (%2),%0"
                     : "=q" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::store_fence(jshort* p, jshort v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jshort>(volatile jshort* p, jshort v) {
   __asm__ volatile (  "xchgw (%2),%0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::store_fence(jint*   p, jint   v) {
+template<>
+inline void OrderAccess::specialized_release_store_fence<jint>  (volatile jint*   p, jint   v) {
   __asm__ volatile (  "xchgl (%2),%0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
 
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) {
 #ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  *p = v; fence();
-#endif // AMD64
-}
-
-// AMD64 copied the bodies for the the signed version. 32bit did this. As long as the
-// compiler does the inlining this is simpler.
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { store_fence((jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { store_fence((jshort*)p, (jshort)v); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { store_fence((jint*)p,   (jint)v);   }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { store_fence((jlong*)p,  (jlong)v);  }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { store_fence((jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
-}
-
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) {
-#ifdef AMD64
-  __asm__ __volatile__ ("xchgq (%2), %0"
-                        : "=r" (v)
-                        : "0" (v), "r" (p)
-                        : "memory");
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
-}
-
-// Must duplicate definitions instead of calling store_fence because we don't want to cast away volatile.
-inline void     OrderAccess::release_store_fence(volatile jbyte*  p, jbyte  v) {
-  __asm__ volatile (  "xchgb (%2),%0"
-                    : "=q" (v)
-                    : "0" (v), "r" (p)
-                    : "memory");
-}
-inline void     OrderAccess::release_store_fence(volatile jshort* p, jshort v) {
-  __asm__ volatile (  "xchgw (%2),%0"
+template<>
+inline void OrderAccess::specialized_release_store_fence<jlong> (volatile jlong*  p, jlong  v) {
+  __asm__ volatile (  "xchgq (%2), %0"
                     : "=r" (v)
                     : "0" (v), "r" (p)
                     : "memory");
 }
-inline void     OrderAccess::release_store_fence(volatile jint*   p, jint   v) {
-  __asm__ volatile (  "xchgl (%2),%0"
-                    : "=r" (v)
-                    : "0" (v), "r" (p)
-                    : "memory");
-}
+#endif // AMD64
 
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store(p, v); fence();
-#endif // AMD64
+template<>
+inline void OrderAccess::specialized_release_store_fence<jfloat> (volatile jfloat*  p, jfloat  v) {
+  release_store_fence((volatile jint*)p, jint_cast(v));
+}
+template<>
+inline void OrderAccess::specialized_release_store_fence<jdouble>(volatile jdouble* p, jdouble v) {
+  release_store_fence((volatile jlong*)p, jlong_cast(v));
 }
 
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { release_store_fence((volatile jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { release_store_fence((volatile jshort*)p, (jshort)v); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { release_store_fence((volatile jint*)p,   (jint)v);   }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store_fence((volatile jlong*)p,  (jlong)v);  }
-
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store_fence((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) {
-#ifdef AMD64
-  __asm__ __volatile__ (  "xchgq (%2), %0"
-                          : "=r" (v)
-                          : "0" (v), "r" (p)
-                          : "memory");
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_LINUX_X86_VM_ORDERACCESS_LINUX_X86_INLINE_HPP
--- a/hotspot/src/os_cpu/linux_zero/vm/orderAccess_linux_zero.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/linux_zero/vm/orderAccess_linux_zero.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 2015, Oracle and/or its affiliates. All rights reserved.
  * Copyright 2007, 2008, 2009 Red Hat, Inc.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -40,8 +40,7 @@
 #define __kernel_dmb (*(__kernel_dmb_t *) 0xffff0fa0)
 
 #define FULL_MEM_BARRIER __kernel_dmb()
-#define READ_MEM_BARRIER __kernel_dmb()
-#define WRITE_MEM_BARRIER __kernel_dmb()
+#define LIGHT_MEM_BARRIER __kernel_dmb()
 
 #else // ARM
 
@@ -49,126 +48,33 @@
 
 #ifdef PPC
 
-#define READ_MEM_BARRIER __asm __volatile ("isync":::"memory")
 #ifdef __NO_LWSYNC__
-#define WRITE_MEM_BARRIER __asm __volatile ("sync":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("sync":::"memory")
 #else
-#define WRITE_MEM_BARRIER __asm __volatile ("lwsync":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("lwsync":::"memory")
 #endif
 
 #else // PPC
 
-#define READ_MEM_BARRIER __asm __volatile ("":::"memory")
-#define WRITE_MEM_BARRIER __asm __volatile ("":::"memory")
+#define LIGHT_MEM_BARRIER __asm __volatile ("":::"memory")
 
 #endif // PPC
 
 #endif // ARM
 
-
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
-
-inline void OrderAccess::acquire() {
-  READ_MEM_BARRIER;
-}
-
-inline void OrderAccess::release() {
-  WRITE_MEM_BARRIER;
-}
-
-inline void OrderAccess::fence() {
-  FULL_MEM_BARRIER;
-}
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { jbyte data = *p; acquire(); return data; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { jshort data = *p; acquire(); return data; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { jint data = *p; acquire(); return data; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) {
-  jlong tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
-inline jubyte    OrderAccess::load_acquire(volatile jubyte*   p) { jubyte data = *p; acquire(); return data; }
-inline jushort   OrderAccess::load_acquire(volatile jushort*  p) { jushort data = *p; acquire(); return data; }
-inline juint     OrderAccess::load_acquire(volatile juint*    p) { juint data = *p; acquire(); return data; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) {
-  julong tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { jfloat data = *p; acquire(); return data; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) {
-  jdouble tmp;
-  os::atomic_copy64(p, &tmp);
-  acquire();
-  return tmp;
-}
+// Note: What is meant by LIGHT_MEM_BARRIER is a barrier which is sufficient
+// to provide TSO semantics, i.e. StoreStore | LoadLoad | LoadStore.
 
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) {
-  intptr_t data = *p;
-  acquire();
-  return data;
-}
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) {
-  void *data = *(void* volatile *)p;
-  acquire();
-  return data;
-}
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) {
-  void *data = *(void* const volatile *)p;
-  acquire();
-  return data;
-}
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v)
-{ release(); os::atomic_copy64(&v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v)
-{ release(); os::atomic_copy64(&v, p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { release(); *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v)
-{ release(); os::atomic_copy64(&v, p); }
+inline void OrderAccess::loadload()   { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::storestore() { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::loadstore()  { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::storeload()  { FULL_MEM_BARRIER;  }
 
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { release(); *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v)
-{ release(); *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { os::atomic_copy64(&v, p); fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { os::atomic_copy64(&v, p); fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { os::atomic_copy64(&v, p); fence(); }
+inline void OrderAccess::acquire()    { LIGHT_MEM_BARRIER; }
+inline void OrderAccess::release()    { LIGHT_MEM_BARRIER; }
 
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; fence(); }
+inline void OrderAccess::fence()      { FULL_MEM_BARRIER;  }
 
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store(p, v); fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { release_store_ptr(p, v); fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { release_store_ptr(p, v); fence(); }
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_LINUX_ZERO_VM_ORDERACCESS_LINUX_ZERO_INLINE_HPP
--- a/hotspot/src/os_cpu/solaris_sparc/vm/orderAccess_solaris_sparc.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/solaris_sparc/vm/orderAccess_solaris_sparc.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -28,107 +28,30 @@
 #include "runtime/atomic.inline.hpp"
 #include "runtime/orderAccess.hpp"
 
+// Compiler version last used for testing: solaris studio 12u3
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
 // Assume TSO.
 
-// In solaris_sparc.il
-extern "C" void _OrderAccess_acquire();
-extern "C" void _OrderAccess_fence();
-
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
-
-#ifdef _GNU_SOURCE
-
-inline void OrderAccess::acquire() {
-  __asm__ volatile ("nop" : : :);
-}
-
-inline void OrderAccess::release() {
-  jint* local_dummy = (jint*)&local_dummy;
-  __asm__ volatile("stw %%g0, [%0]" : : "r" (local_dummy) : "memory");
-}
-
-inline void OrderAccess::fence() {
-  __asm__ volatile ("membar  #StoreLoad" : : :);
-}
-
-#else  // _GNU_SOURCE
-
-inline void OrderAccess::acquire() {
-  _OrderAccess_acquire();
-}
-
-inline void OrderAccess::release() {
-  // Avoid hitting the same cache-line from
-  // different threads.
-  volatile jint local_dummy = 0;
-}
-
-inline void OrderAccess::fence() {
-  _OrderAccess_fence();
+// A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
+inline void compiler_barrier() {
+  __asm__ volatile ("" : : : "memory");
 }
 
-#endif // _GNU_SOURCE
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return *p; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return *p; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { return *p; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return Atomic::load(p); }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return *p; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return *p; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { return *p; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return Atomic::load((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return *p; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return *p; }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return *p; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return *(void* volatile *)p; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return *(void* const volatile *)p; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { Atomic::store(v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { Atomic::store((jlong)v, (volatile jlong*)p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { *p = v; }
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
+inline void OrderAccess::storeload()  { fence();            }
 
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; fence(); }
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; fence(); }
+inline void OrderAccess::fence() {
+  __asm__ volatile ("membar  #StoreLoad" : : : "memory");
+}
 
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { *p = v; fence(); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { *(void* volatile *)p = v; fence(); }
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_SOLARIS_SPARC_VM_ORDERACCESS_SOLARIS_SPARC_INLINE_HPP
--- a/hotspot/src/os_cpu/solaris_sparc/vm/solaris_sparc.il	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/solaris_sparc/vm/solaris_sparc.il	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 //
-// Copyright (c) 2002, 2013, Oracle and/or its affiliates. All rights reserved.
+// Copyright (c) 2002, 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
@@ -19,7 +19,7 @@
 // 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.
-//  
+//
 //
 
   // Get the raw thread ID from %g7
@@ -35,11 +35,11 @@
   // Clear SPARC fprs.FEF DU and DL bits --
   // allows the kernel to avoid saving FPU state at context-switch time.
   // Use for state-transition points (into _thread_blocked) or when
-  // parking. 
-      
+  // parking.
+
        .inline _mark_fpu_nosave, 0
        .volatile
-       wr   %g0, 0, %fprs       
+       wr   %g0, 0, %fprs
        .nonvolatile
        .end
 
@@ -85,7 +85,7 @@
 
 
   // Support for jint Atomic::cmpxchg(jint           exchange_value,
-  //                                  volatile jint* dest, 
+  //                                  volatile jint* dest,
   //                                  jint           compare_value)
   //
   // Arguments:
@@ -103,8 +103,8 @@
         .end
 
 
-  // Support for intptr_t Atomic::cmpxchg_ptr(intptr_t           exchange_value, 
-  //                                          volatile intptr_t* dest, 
+  // Support for intptr_t Atomic::cmpxchg_ptr(intptr_t           exchange_value,
+  //                                          volatile intptr_t* dest,
   //                                          intptr_t           compare_value)
   //
   // 64-bit
@@ -124,8 +124,8 @@
         .end
 
 
-  // Support for jlong Atomic::cmpxchg(jlong           exchange_value, 
-  //                                   volatile jlong* dest, 
+  // Support for jlong Atomic::cmpxchg(jlong           exchange_value,
+  //                                   volatile jlong* dest,
   //                                   jlong           compare_value)
   //
   // 32-bit calling conventions
@@ -221,27 +221,6 @@
         .end
 
 
-  // Support for void OrderAccess::acquire()
-  // The method is intentionally empty.  
-  // It exists for the sole purpose of generating
-  // a C/C++ sequence point over which the compiler won't 
-  // reorder code.
-
-        .inline _OrderAccess_acquire,0
-        .volatile
-        .nonvolatile
-        .end
-
-
-  // Support for void OrderAccess::fence()
-
-        .inline _OrderAccess_fence,0
-        .volatile
-        membar  #StoreLoad
-        .nonvolatile
-        .end
-
-
   // Support for void Prefetch::read(void *loc, intx interval)
   //
   // Prefetch for several reads.
--- a/hotspot/src/os_cpu/solaris_x86/vm/orderAccess_solaris_x86.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/solaris_x86/vm/orderAccess_solaris_x86.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -29,110 +29,35 @@
 #include "runtime/orderAccess.hpp"
 #include "runtime/os.hpp"
 
+// Compiler version last used for testing: solaris studio 12u3
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
-// For Sun Studio - implementation is in solaris_i486.il.
-// For gcc - implementation is just below.
-extern "C" void _OrderAccess_acquire();
-extern "C" void _OrderAccess_fence();
-
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
-inline void OrderAccess::storeload()  { fence(); }
-
-inline void OrderAccess::acquire() {
-  _OrderAccess_acquire();
-
+// A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
+inline void compiler_barrier() {
+  __asm__ volatile ("" : : : "memory");
 }
 
-inline void OrderAccess::release() {
-  // Avoid hitting the same cache-line from
-  // different threads.
-  volatile jint local_dummy = 0;
-}
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
+inline void OrderAccess::storeload()  { fence();            }
+
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
 inline void OrderAccess::fence() {
   if (os::is_MP()) {
-    _OrderAccess_fence();
+#ifdef AMD64
+    __asm__ volatile ("lock; addl $0,0(%%rsp)" : : : "cc", "memory");
+#else
+    __asm__ volatile ("lock; addl $0,0(%%esp)" : : : "cc", "memory");
+#endif
   }
-}
-
-#ifdef _GNU_SOURCE
-
-extern "C" {
-  inline void _OrderAccess_acquire() {
-    volatile intptr_t local_dummy;
-#ifdef AMD64
-    __asm__ volatile ("movq 0(%%rsp), %0" : "=r" (local_dummy) : : "memory");
-#else
-    __asm__ volatile ("movl 0(%%esp),%0" : "=r" (local_dummy) : : "memory");
-#endif // AMD64
-  }
-  inline void _OrderAccess_fence() {
-    // Always use locked addl since mfence is sometimes expensive
-    __asm__ volatile ("lock; addl $0,0(%%esp)" : : : "cc", "memory");
-  }
-
+  compiler_barrier();
 }
 
-#endif // GNU_SOURCE
-
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return *p; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return *p; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { return *p; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return Atomic::load(p); }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return *p; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return *p; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { return *p; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return Atomic::load((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return *p; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return jdouble_cast(Atomic::load((volatile jlong*)p)); }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return *p; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return *(void* volatile *)p; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return *(void* const volatile *)p; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { Atomic::store(v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { Atomic::store((jlong)v, (volatile jlong*)p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { release_store((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; fence(); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) { *p = v; fence(); }
-
-inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { release_store(p, v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store((jlong *)p, (jlong)v); fence(); }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store_fence((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { *(void* volatile *)p = v; fence(); }
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_SOLARIS_X86_VM_ORDERACCESS_SOLARIS_X86_INLINE_HPP
--- a/hotspot/src/os_cpu/solaris_x86/vm/solaris_x86_32.il	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/solaris_x86/vm/solaris_x86_32.il	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 //
-// Copyright (c) 2003, 2012, Oracle and/or its affiliates. All rights reserved.
+// Copyright (c) 2003, 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
@@ -19,7 +19,7 @@
 // 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.
-//  
+//
 //
 
 
@@ -34,19 +34,19 @@
 
   // Get the raw thread ID from %gs:0
       .inline _raw_thread_id,0
-      movl     %gs:0, %eax 
+      movl     %gs:0, %eax
       .end
 
   // Get current sp
       .inline _get_current_sp,0
       .volatile
-      movl     %esp, %eax 
+      movl     %esp, %eax
       .end
 
   // Get current fp
       .inline _get_current_fp,0
       .volatile
-      movl     %ebp, %eax 
+      movl     %ebp, %eax
       .end
 
   // Support for os::rdtsc()
@@ -76,8 +76,8 @@
       xchgl    (%ecx), %eax
       .end
 
-  // Support for jbyte Atomic::cmpxchg(jbyte exchange_value, 
-  //                                   volatile jbyte *dest, 
+  // Support for jbyte Atomic::cmpxchg(jbyte exchange_value,
+  //                                   volatile jbyte *dest,
   //                                   jbyte compare_value)
   // An additional bool (os::is_MP()) is passed as the last argument.
       .inline _Atomic_cmpxchg_byte,4
@@ -93,8 +93,8 @@
 2:
       .end
 
-  // Support for jint Atomic::cmpxchg(jint exchange_value, 
-  //                                  volatile jint *dest, 
+  // Support for jint Atomic::cmpxchg(jint exchange_value,
+  //                                  volatile jint *dest,
   //                                  jint compare_value)
   // An additional bool (os::is_MP()) is passed as the last argument.
       .inline _Atomic_cmpxchg,4
@@ -141,17 +141,6 @@
       fistpll   (%eax)
       .end
 
-  // Support for OrderAccess::acquire()
-      .inline _OrderAccess_acquire,0
-      movl     0(%esp), %eax
-      .end
-
-  // Support for OrderAccess::fence()
-      .inline _OrderAccess_fence,0
-      lock
-      addl     $0, (%esp)
-      .end
-
   // Support for u2 Bytes::swap_u2(u2 x)
       .inline _raw_swap_u2,1
       movl     0(%esp), %eax
--- a/hotspot/src/os_cpu/solaris_x86/vm/solaris_x86_64.il	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/solaris_x86/vm/solaris_x86_64.il	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 //
-// Copyright (c) 2004, 2012, Oracle and/or its affiliates. All rights reserved.
+// Copyright (c) 2004, 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
@@ -19,7 +19,7 @@
 // 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.
-//  
+//
 //
 
   // The argument size of each inline directive is ignored by the compiler
@@ -27,19 +27,19 @@
 
   // Get the raw thread ID from %gs:0
       .inline _raw_thread_id,0
-      movq     %fs:0, %rax 
+      movq     %fs:0, %rax
       .end
 
   // Get current sp
       .inline _get_current_sp,0
       .volatile
-      movq     %rsp, %rax 
+      movq     %rsp, %rax
       .end
 
   // Get current fp
       .inline _get_current_fp,0
       .volatile
-      movq     %rbp, %rax 
+      movq     %rbp, %rax
       .end
 
   // Support for os::rdtsc()
@@ -77,8 +77,8 @@
       movq     %rdi, %rax
       .end
 
-  // Support for jbyte Atomic::cmpxchg(jbyte exchange_value, 
-  //                                   volatile jbyte *dest, 
+  // Support for jbyte Atomic::cmpxchg(jbyte exchange_value,
+  //                                   volatile jbyte *dest,
   //                                   jbyte compare_value)
       .inline _Atomic_cmpxchg_byte,3
       movb     %dl, %al      // compare_value
@@ -86,8 +86,8 @@
       cmpxchgb %dil, (%rsi)
       .end
 
-  // Support for jint Atomic::cmpxchg(jint exchange_value, 
-  //                                  volatile jint *dest, 
+  // Support for jint Atomic::cmpxchg(jint exchange_value,
+  //                                  volatile jint *dest,
   //                                  jint compare_value)
       .inline _Atomic_cmpxchg,3
       movl     %edx, %eax      // compare_value
@@ -104,17 +104,6 @@
       cmpxchgq %rdi, (%rsi)
       .end
 
-  // Support for OrderAccess::acquire()
-      .inline _OrderAccess_acquire,0
-      movl     0(%rsp), %eax
-      .end
-
-  // Support for OrderAccess::fence()
-      .inline _OrderAccess_fence,0
-      lock
-      addl     $0, (%rsp)
-      .end
-
   // Support for u2 Bytes::swap_u2(u2 x)
       .inline _raw_swap_u2,1
       movw     %di, %ax
--- a/hotspot/src/os_cpu/windows_x86/vm/orderAccess_windows_x86.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/os_cpu/windows_x86/vm/orderAccess_windows_x86.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -25,29 +25,39 @@
 #ifndef OS_CPU_WINDOWS_X86_VM_ORDERACCESS_WINDOWS_X86_INLINE_HPP
 #define OS_CPU_WINDOWS_X86_VM_ORDERACCESS_WINDOWS_X86_INLINE_HPP
 
+#include <intrin.h>
 #include "runtime/atomic.inline.hpp"
 #include "runtime/orderAccess.hpp"
 #include "runtime/os.hpp"
 
+// Compiler version last used for testing: Microsoft Visual Studio 2010
+// Please update this information when this file changes
+
 // Implementation of class OrderAccess.
 
-inline void OrderAccess::loadload()   { acquire(); }
-inline void OrderAccess::storestore() { release(); }
-inline void OrderAccess::loadstore()  { acquire(); }
+// A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions
+inline void compiler_barrier() {
+  _ReadWriteBarrier();
+}
+
+// Note that in MSVC, volatile memory accesses are explicitly
+// guaranteed to have acquire release semantics (w.r.t. compiler
+// reordering) and therefore does not even need a compiler barrier
+// for normal acquire release accesses. And all generalized
+// bound calls like release_store go through OrderAccess::load
+// and OrderAccess::store which do volatile memory accesses.
+template<> inline void ScopedFence<X_ACQUIRE>::postfix()       { }
+template<> inline void ScopedFence<RELEASE_X>::prefix()        { }
+template<> inline void ScopedFence<RELEASE_X_FENCE>::prefix()  { }
+template<> inline void ScopedFence<RELEASE_X_FENCE>::postfix() { OrderAccess::fence(); }
+
+inline void OrderAccess::loadload()   { compiler_barrier(); }
+inline void OrderAccess::storestore() { compiler_barrier(); }
+inline void OrderAccess::loadstore()  { compiler_barrier(); }
 inline void OrderAccess::storeload()  { fence(); }
 
-inline void OrderAccess::acquire() {
-#ifndef AMD64
-  __asm {
-    mov eax, dword ptr [esp];
-  }
-#endif // !AMD64
-}
-
-inline void OrderAccess::release() {
-  // A volatile store has release semantics.
-  volatile jint local_dummy = 0;
-}
+inline void OrderAccess::acquire()    { compiler_barrier(); }
+inline void OrderAccess::release()    { compiler_barrier(); }
 
 inline void OrderAccess::fence() {
 #ifdef AMD64
@@ -59,157 +69,47 @@
     }
   }
 #endif // AMD64
+  compiler_barrier();
 }
 
-inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return *p; }
-inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return *p; }
-inline jint     OrderAccess::load_acquire(volatile jint*    p) { return *p; }
-inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return Atomic::load(p); }
-inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return *p; }
-inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return *p; }
-inline juint    OrderAccess::load_acquire(volatile juint*   p) { return *p; }
-inline julong   OrderAccess::load_acquire(volatile julong*  p) { return Atomic::load((volatile jlong*)p); }
-inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return *p; }
-inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return jdouble_cast(Atomic::load((volatile jlong*)p)); }
-
-inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return *p; }
-inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return *(void* volatile *)p; }
-inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return *(void* const volatile *)p; }
-
-inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { Atomic::store(v, p); }
-inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { *p = v; }
-inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { *p = v; }
-inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { Atomic::store((jlong)v, (volatile jlong*)p); }
-inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { *p = v; }
-inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { release_store((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { *p = v; }
-inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { *(void* volatile *)p = v; }
-
-inline void     OrderAccess::store_fence(jbyte*  p, jbyte  v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
+#ifndef AMD64
+template<>
+inline void OrderAccess::specialized_release_store_fence<jbyte> (volatile jbyte*  p, jbyte  v) {
   __asm {
     mov edx, p;
     mov al, v;
     xchg al, byte ptr [edx];
   }
-#endif // AMD64
 }
 
-inline void     OrderAccess::store_fence(jshort* p, jshort v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
+template<>
+inline void OrderAccess::specialized_release_store_fence<jshort>(volatile jshort* p, jshort v) {
   __asm {
     mov edx, p;
     mov ax, v;
     xchg ax, word ptr [edx];
   }
-#endif // AMD64
 }
 
-inline void     OrderAccess::store_fence(jint*   p, jint   v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
+template<>
+inline void OrderAccess::specialized_release_store_fence<jint>  (volatile jint*   p, jint   v) {
   __asm {
     mov edx, p;
     mov eax, v;
     xchg eax, dword ptr [edx];
   }
+}
 #endif // AMD64
-}
-
-inline void     OrderAccess::store_fence(jlong*   p, jlong   v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jubyte*  p, jubyte  v) { store_fence((jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::store_fence(jushort* p, jushort v) { store_fence((jshort*)p, (jshort)v); }
-inline void     OrderAccess::store_fence(juint*   p, juint   v) { store_fence((jint*)p,   (jint)v);   }
-inline void     OrderAccess::store_fence(julong*  p, julong  v) { store_fence((jlong*)p,  (jlong)v);  }
-inline void     OrderAccess::store_fence(jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::store_fence(jdouble* p, jdouble v) { *p = v; fence(); }
-
-inline void     OrderAccess::store_ptr_fence(intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
-}
 
-inline void     OrderAccess::store_ptr_fence(void**    p, void*    v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  store_fence((jint*)p, (jint)v);
-#endif // AMD64
+template<>
+inline void OrderAccess::specialized_release_store_fence<jfloat>(volatile jfloat*  p, jfloat  v) {
+    release_store_fence((volatile jint*)p, jint_cast(v));
 }
-
-// Must duplicate definitions instead of calling store_fence because we don't want to cast away volatile.
-inline void     OrderAccess::release_store_fence(volatile jbyte*  p, jbyte  v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  __asm {
-    mov edx, p;
-    mov al, v;
-    xchg al, byte ptr [edx];
-  }
-#endif // AMD64
+template<>
+inline void OrderAccess::specialized_release_store_fence<jdouble>(volatile jdouble* p, jdouble v) {
+    release_store_fence((volatile jlong*)p, jlong_cast(v));
 }
 
-inline void     OrderAccess::release_store_fence(volatile jshort* p, jshort v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  __asm {
-    mov edx, p;
-    mov ax, v;
-    xchg ax, word ptr [edx];
-  }
-#endif // AMD64
-}
-
-inline void     OrderAccess::release_store_fence(volatile jint*   p, jint   v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  __asm {
-    mov edx, p;
-    mov eax, v;
-    xchg eax, dword ptr [edx];
-  }
-#endif // AMD64
-}
-
-inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { release_store(p, v); fence(); }
-
-inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { release_store_fence((volatile jbyte*)p,  (jbyte)v);  }
-inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { release_store_fence((volatile jshort*)p, (jshort)v); }
-inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { release_store_fence((volatile jint*)p,   (jint)v);   }
-inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { release_store_fence((volatile jlong*)p,  (jlong)v);  }
-inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { *p = v; fence(); }
-inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { release_store_fence((volatile jlong*)p, jlong_cast(v)); }
-
-inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) {
-#ifdef AMD64
-  *p = v; fence();
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
-
-inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) {
-#ifdef AMD64
-  *(void* volatile *)p = v; fence();
-#else
-  release_store_fence((volatile jint*)p, (jint)v);
-#endif // AMD64
-}
+#define VM_HAS_GENERALIZED_ORDER_ACCESS 1
 
 #endif // OS_CPU_WINDOWS_X86_VM_ORDERACCESS_WINDOWS_X86_INLINE_HPP
--- a/hotspot/src/share/vm/runtime/arguments.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/runtime/arguments.cpp	Thu Mar 12 16:10:58 2015 +0100
@@ -321,6 +321,8 @@
   { "UseFastEmptyMethods",           JDK_Version::jdk(9), JDK_Version::jdk(10) },
 #endif // ZERO
   { "UseCompilerSafepoints",         JDK_Version::jdk(9), JDK_Version::jdk(10) },
+  { "AdaptiveSizePausePolicy",       JDK_Version::jdk(9), JDK_Version::jdk(10) },
+  { "ParallelGCRetainPLAB",          JDK_Version::jdk(9), JDK_Version::jdk(10) },
   { NULL, JDK_Version(0), JDK_Version(0) }
 };
 
--- a/hotspot/src/share/vm/runtime/globals.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/runtime/globals.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1080,9 +1080,6 @@
   notproduct(bool, ProfilerCheckIntervals, false,                           \
           "Collect and print information on spacing of profiler ticks")     \
                                                                             \
-  develop(bool, PrintJVMWarnings, false,                                    \
-          "Print warnings for unimplemented JVM functions")                 \
-                                                                            \
   product(bool, PrintWarnings, true,                                        \
           "Print JVM warnings to output stream")                            \
                                                                             \
@@ -1207,10 +1204,6 @@
           "Use pthread-based instead of libthread-based synchronization "   \
           "(SPARC only)")                                                   \
                                                                             \
-  product(bool, AdjustConcurrency, false,                                   \
-          "Call thr_setconcurrency at thread creation time to avoid "       \
-          "LWP starvation on MP systems (for Solaris Only)")                \
-                                                                            \
   product(bool, ReduceSignalUsage, false,                                   \
           "Reduce the use of OS signals in Java and/or the VM")             \
                                                                             \
@@ -1557,11 +1550,6 @@
   product(uintx, ParallelGCBufferWastePct, 10,                              \
           "Wasted fraction of parallel allocation buffer")                  \
                                                                             \
-  diagnostic(bool, ParallelGCRetainPLAB, false,                             \
-          "Retain parallel allocation buffers across scavenges; "           \
-          "it is disabled because this currently conflicts with "           \
-          "parallel card scanning under certain conditions.")               \
-                                                                            \
   product(uintx, TargetPLABWastePct, 10,                                    \
           "Target wasted space in last buffer as percent of overall "       \
           "allocation")                                                     \
@@ -2101,9 +2089,6 @@
   product(uintx, AdaptiveSizeThroughPutPolicy, 0,                           \
           "Policy for changing generation size for throughput goals")       \
                                                                             \
-  product(uintx, AdaptiveSizePausePolicy, 0,                                \
-          "Policy for changing generation size for pause goals")            \
-                                                                            \
   develop(bool, PSAdjustTenuredGenForMinorPause, false,                     \
           "Adjust tenured generation to achieve a minor pause goal")        \
                                                                             \
--- a/hotspot/src/share/vm/runtime/orderAccess.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/runtime/orderAccess.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2003, 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
@@ -29,11 +29,7 @@
 
 //                Memory Access Ordering Model
 //
-// This interface is based on the JSR-133 Cookbook for Compiler Writers
-// and on the IA64 memory model.  It is the dynamic equivalent of the
-// C/C++ volatile specifier.  I.e., volatility restricts compile-time
-// memory access reordering in a way similar to what we want to occur
-// at runtime.
+// This interface is based on the JSR-133 Cookbook for Compiler Writers.
 //
 // In the following, the terms 'previous', 'subsequent', 'before',
 // 'after', 'preceding' and 'succeeding' refer to program order.  The
@@ -41,7 +37,6 @@
 // relative to program order, while 'up' and 'above' refer to backward
 // motion.
 //
-//
 // We define four primitive memory barrier operations.
 //
 // LoadLoad:   Load1(s); LoadLoad; Load2
@@ -69,86 +64,88 @@
 // operations.  Stores before Store1 may *not* float below Load2 and any
 // subsequent load operations.
 //
+// We define two further barriers: acquire and release.
 //
-// We define two further operations, 'release' and 'acquire'.  They are
-// mirror images of each other.
+// Conceptually, acquire/release semantics form unidirectional and
+// asynchronous barriers w.r.t. a synchronizing load(X) and store(X) pair.
+// They should always be used in pairs to publish (release store) and
+// access (load acquire) some implicitly understood shared data between
+// threads in a relatively cheap fashion not requiring storeload. If not
+// used in such a pair, it is advised to use a membar instead:
+// acquire/release only make sense as pairs.
+//
+// T1: access_shared_data
+// T1: ]release
+// T1: (...)
+// T1: store(X)
 //
-// Execution by a processor of release makes the effect of all memory
-// accesses issued by it previous to the release visible to all
-// processors *before* the release completes.  The effect of subsequent
-// memory accesses issued by it *may* be made visible *before* the
-// release.  I.e., subsequent memory accesses may float above the
-// release, but prior ones may not float below it.
+// T2: load(X)
+// T2: (...)
+// T2: acquire[
+// T2: access_shared_data
+//
+// It is guaranteed that if T2: load(X) synchronizes with (observes the
+// value written by) T1: store(X), then the memory accesses before the T1:
+// ]release happen before the memory accesses after the T2: acquire[.
+//
+// Total Store Order (TSO) machines can be seen as machines issuing a
+// release store for each store and a load acquire for each load. Therefore
+// there is an inherent resemblence between TSO and acquire/release
+// semantics. TSO can be seen as an abstract machine where loads are
+// executed immediately when encountered (hence loadload reordering not
+// happening) but enqueues stores in a FIFO queue
+// for asynchronous serialization (neither storestore or loadstore
+// reordering happening). The only reordering happening is storeload due to
+// the queue asynchronously serializing stores (yet in order).
 //
-// Execution by a processor of acquire makes the effect of all memory
-// accesses issued by it subsequent to the acquire visible to all
-// processors *after* the acquire completes.  The effect of prior memory
-// accesses issued by it *may* be made visible *after* the acquire.
-// I.e., prior memory accesses may float below the acquire, but
-// subsequent ones may not float above it.
+// Acquire/release semantics essentially exploits this asynchronicity: when
+// the load(X) acquire[ observes the store of ]release store(X), the
+// accesses before the release must have happened before the accesses after
+// acquire.
+//
+// The API offers both stand-alone acquire() and release() as well as bound
+// load_acquire() and release_store(). It is guaranteed that these are
+// semantically equivalent w.r.t. the defined model. However, since
+// stand-alone acquire()/release() does not know which previous
+// load/subsequent store is considered the synchronizing load/store, they
+// may be more conservative in implementations. We advise using the bound
+// variants whenever possible.
+//
+// Finally, we define a "fence" operation, as a bidirectional barrier.
+// It guarantees that any memory access preceding the fence is not
+// reordered w.r.t. any memory accesses subsequent to the fence in program
+// order. This may be used to prevent sequences of loads from floating up
+// above sequences of stores.
 //
-// Finally, we define a 'fence' operation, which conceptually is a
-// release combined with an acquire.  In the real world these operations
-// require one or more machine instructions which can float above and
-// below the release or acquire, so we usually can't just issue the
-// release-acquire back-to-back.  All machines we know of implement some
-// sort of memory fence instruction.
+// The following table shows the implementations on some architectures:
+//
+//                       Constraint     x86          sparc TSO          ppc
+// ---------------------------------------------------------------------------
+// fence                 LoadStore  |   lock         membar #StoreLoad  sync
+//                       StoreStore |   addl 0,(sp)
+//                       LoadLoad   |
+//                       StoreLoad
+//
+// release               LoadStore  |                                   lwsync
+//                       StoreStore
+//
+// acquire               LoadLoad   |                                   lwsync
+//                       LoadStore
+//
+// release_store                        <store>      <store>            lwsync
+//                                                                      <store>
+//
+// release_store_fence                  xchg         <store>            lwsync
+//                                                   membar #StoreLoad  <store>
+//                                                                      sync
 //
 //
-// The standalone implementations of release and acquire need an associated
-// dummy volatile store or load respectively.  To avoid redundant operations,
-// we can define the composite operators: 'release_store', 'store_fence' and
-// 'load_acquire'.  Here's a summary of the machine instructions corresponding
-// to each operation.
-//
-//               sparc RMO             ia64             x86
-// ---------------------------------------------------------------------
-// fence         membar #LoadStore |   mf               lock addl 0,(sp)
-//                      #StoreStore |
-//                      #LoadLoad |
-//                      #StoreLoad
-//
-// release       membar #LoadStore |   st.rel [sp]=r0   movl $0,<dummy>
-//                      #StoreStore
-//               st %g0,[]
-//
-// acquire       ld [%sp],%g0          ld.acq <r>=[sp]  movl (sp),<r>
-//               membar #LoadLoad |
-//                      #LoadStore
-//
-// release_store membar #LoadStore |   st.rel           <store>
-//                      #StoreStore
-//               st
+// load_acquire                         <load>       <load>             <load>
+//                                                                      lwsync
 //
-// store_fence   st                    st               lock xchg
-//               fence                 mf
-//
-// load_acquire  ld                    ld.acq           <load>
-//               membar #LoadLoad |
-//                      #LoadStore
-//
-// Using only release_store and load_acquire, we can implement the
-// following ordered sequences.
-//
-// 1. load, load   == load_acquire,  load
-//                 or load_acquire,  load_acquire
-// 2. load, store  == load,          release_store
-//                 or load_acquire,  store
-//                 or load_acquire,  release_store
-// 3. store, store == store,         release_store
-//                 or release_store, release_store
-//
-// These require no membar instructions for sparc-TSO and no extra
-// instructions for ia64.
-//
-// Ordering a load relative to preceding stores requires a store_fence,
+// Ordering a load relative to preceding stores requires a StoreLoad,
 // which implies a membar #StoreLoad between the store and load under
-// sparc-TSO.  A fence is required by ia64.  On x86, we use locked xchg.
-//
-// 4. store, load  == store_fence, load
-//
-// Use store_fence to make sure all stores done in an 'interesting'
-// region are made visible prior to both subsequent loads and stores.
+// sparc-TSO. On x86, we use explicitly locked add.
 //
 // Conventional usage is to issue a load_acquire for ordered loads.  Use
 // release_store for ordered stores when you care only that prior stores
@@ -157,27 +154,19 @@
 // release_store_fence to update values like the thread state, where we
 // don't want the current thread to continue until all our prior memory
 // accesses (including the new thread state) are visible to other threads.
-//
+// This is equivalent to the volatile semantics of the Java Memory Model.
 //
-//                C++ Volatility
+//                    C++ Volatile Semantics
 //
-// C++ guarantees ordering at operations termed 'sequence points' (defined
-// to be volatile accesses and calls to library I/O functions).  'Side
-// effects' (defined as volatile accesses, calls to library I/O functions
-// and object modification) previous to a sequence point must be visible
-// at that sequence point.  See the C++ standard, section 1.9, titled
-// "Program Execution".  This means that all barrier implementations,
-// including standalone loadload, storestore, loadstore, storeload, acquire
-// and release must include a sequence point, usually via a volatile memory
-// access.  Other ways to guarantee a sequence point are, e.g., use of
-// indirect calls and linux's __asm__ volatile.
-// Note: as of 6973570, we have replaced the originally static "dummy" field
-// (see above) by a volatile store to the stack. All of the versions of the
-// compilers that we currently use (SunStudio, gcc and VC++) respect the
-// semantics of volatile here. If you build HotSpot using other
-// compilers, you may need to verify that no compiler reordering occurs
-// across the sequence point represented by the volatile access.
-//
+// C++ volatile semantics prevent compiler re-ordering between
+// volatile memory accesses. However, reordering between non-volatile
+// and volatile memory accesses is in general undefined. For compiler
+// reordering constraints taking non-volatile memory accesses into
+// consideration, a compiler barrier has to be used instead.  Some
+// compiler implementations may choose to enforce additional
+// constraints beyond those required by the language. Note also that
+// both volatile semantics and compiler barrier do not prevent
+// hardware reordering.
 //
 //                os::is_MP Considered Redundant
 //
@@ -240,8 +229,32 @@
 // order.  If their implementations change such that these assumptions
 // are violated, a whole lot of code will break.
 
+enum ScopedFenceType {
+    X_ACQUIRE
+  , RELEASE_X
+  , RELEASE_X_FENCE
+};
+
+template <ScopedFenceType T>
+class ScopedFenceGeneral: public StackObj {
+ public:
+  void prefix() {}
+  void postfix() {}
+};
+
+template <ScopedFenceType T>
+class ScopedFence : public ScopedFenceGeneral<T> {
+  void *const _field;
+ public:
+  ScopedFence(void *const field) : _field(field) { prefix(); }
+  ~ScopedFence() { postfix(); }
+  void prefix() { ScopedFenceGeneral<T>::prefix(); }
+  void postfix() { ScopedFenceGeneral<T>::postfix(); }
+};
+
 class OrderAccess : AllStatic {
  public:
+  // barriers
   static void     loadload();
   static void     storestore();
   static void     loadstore();
@@ -280,20 +293,6 @@
   static void     release_store_ptr(volatile intptr_t* p, intptr_t v);
   static void     release_store_ptr(volatile void*     p, void*    v);
 
-  static void     store_fence(jbyte*   p, jbyte   v);
-  static void     store_fence(jshort*  p, jshort  v);
-  static void     store_fence(jint*    p, jint    v);
-  static void     store_fence(jlong*   p, jlong   v);
-  static void     store_fence(jubyte*  p, jubyte  v);
-  static void     store_fence(jushort* p, jushort v);
-  static void     store_fence(juint*   p, juint   v);
-  static void     store_fence(julong*  p, julong  v);
-  static void     store_fence(jfloat*  p, jfloat  v);
-  static void     store_fence(jdouble* p, jdouble v);
-
-  static void     store_ptr_fence(intptr_t* p, intptr_t v);
-  static void     store_ptr_fence(void**    p, void*    v);
-
   static void     release_store_fence(volatile jbyte*   p, jbyte   v);
   static void     release_store_fence(volatile jshort*  p, jshort  v);
   static void     release_store_fence(volatile jint*    p, jint    v);
@@ -313,6 +312,47 @@
   // routine if it exists, It should only be used by platforms that
   // don't have another way to do the inline assembly.
   static void StubRoutines_fence();
+
+  // Give platforms a variation point to specialize.
+  template<typename T> static T    specialized_load_acquire       (volatile T* p     );
+  template<typename T> static void specialized_release_store      (volatile T* p, T v);
+  template<typename T> static void specialized_release_store_fence(volatile T* p, T v);
+
+  template<typename FieldType, ScopedFenceType FenceType>
+  static void ordered_store(volatile FieldType* p, FieldType v);
+
+  template<typename FieldType, ScopedFenceType FenceType>
+  static FieldType ordered_load(volatile FieldType* p);
+
+  static void    store(volatile jbyte*   p, jbyte   v);
+  static void    store(volatile jshort*  p, jshort  v);
+  static void    store(volatile jint*    p, jint    v);
+  static void    store(volatile jlong*   p, jlong   v);
+  static void    store(volatile jdouble* p, jdouble v);
+  static void    store(volatile jfloat*  p, jfloat  v);
+
+  static jbyte   load (volatile jbyte*   p);
+  static jshort  load (volatile jshort*  p);
+  static jint    load (volatile jint*    p);
+  static jlong   load (volatile jlong*   p);
+  static jdouble load (volatile jdouble* p);
+  static jfloat  load (volatile jfloat*  p);
+
+  // The following store_fence methods are deprecated and will be removed
+  // when all repos conform to the new generalized OrderAccess.
+  static void    store_fence(jbyte*   p, jbyte   v);
+  static void    store_fence(jshort*  p, jshort  v);
+  static void    store_fence(jint*    p, jint    v);
+  static void    store_fence(jlong*   p, jlong   v);
+  static void    store_fence(jubyte*  p, jubyte  v);
+  static void    store_fence(jushort* p, jushort v);
+  static void    store_fence(juint*   p, juint   v);
+  static void    store_fence(julong*  p, julong  v);
+  static void    store_fence(jfloat*  p, jfloat  v);
+  static void    store_fence(jdouble* p, jdouble v);
+
+  static void    store_ptr_fence(intptr_t* p, intptr_t v);
+  static void    store_ptr_fence(void**    p, void*    v);
 };
 
 #endif // SHARE_VM_RUNTIME_ORDERACCESS_HPP
--- a/hotspot/src/share/vm/runtime/orderAccess.inline.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/runtime/orderAccess.inline.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2014, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2014, 2015, Oracle and/or its affiliates. All rights reserved.
  * Copyright 2014 SAP AG. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
@@ -26,6 +26,7 @@
 #ifndef SHARE_VM_RUNTIME_ORDERACCESS_INLINE_HPP
 #define SHARE_VM_RUNTIME_ORDERACCESS_INLINE_HPP
 
+#include "runtime/atomic.inline.hpp"
 #include "runtime/orderAccess.hpp"
 
 // Linux
@@ -74,4 +75,92 @@
 # include "orderAccess_bsd_zero.inline.hpp"
 #endif
 
+#ifdef VM_HAS_GENERALIZED_ORDER_ACCESS
+
+template<> inline void ScopedFenceGeneral<X_ACQUIRE>::postfix()       { OrderAccess::acquire(); }
+template<> inline void ScopedFenceGeneral<RELEASE_X>::prefix()        { OrderAccess::release(); }
+template<> inline void ScopedFenceGeneral<RELEASE_X_FENCE>::prefix()  { OrderAccess::release(); }
+template<> inline void ScopedFenceGeneral<RELEASE_X_FENCE>::postfix() { OrderAccess::fence();   }
+
+
+template <typename FieldType, ScopedFenceType FenceType>
+inline void OrderAccess::ordered_store(volatile FieldType* p, FieldType v) {
+  ScopedFence<FenceType> f((void*)p);
+  store(p, v);
+}
+
+template <typename FieldType, ScopedFenceType FenceType>
+inline FieldType OrderAccess::ordered_load(volatile FieldType* p) {
+  ScopedFence<FenceType> f((void*)p);
+  return load(p);
+}
+
+inline jbyte    OrderAccess::load_acquire(volatile jbyte*   p) { return specialized_load_acquire(p); }
+inline jshort   OrderAccess::load_acquire(volatile jshort*  p) { return specialized_load_acquire(p); }
+inline jint     OrderAccess::load_acquire(volatile jint*    p) { return specialized_load_acquire(p); }
+inline jlong    OrderAccess::load_acquire(volatile jlong*   p) { return specialized_load_acquire(p); }
+inline jfloat   OrderAccess::load_acquire(volatile jfloat*  p) { return specialized_load_acquire(p); }
+inline jdouble  OrderAccess::load_acquire(volatile jdouble* p) { return specialized_load_acquire(p); }
+inline jubyte   OrderAccess::load_acquire(volatile jubyte*  p) { return (jubyte) specialized_load_acquire((volatile jbyte*)p);  }
+inline jushort  OrderAccess::load_acquire(volatile jushort* p) { return (jushort)specialized_load_acquire((volatile jshort*)p); }
+inline juint    OrderAccess::load_acquire(volatile juint*   p) { return (juint)  specialized_load_acquire((volatile jint*)p);   }
+inline julong   OrderAccess::load_acquire(volatile julong*  p) { return (julong) specialized_load_acquire((volatile jlong*)p);  }
+
+inline intptr_t OrderAccess::load_ptr_acquire(volatile intptr_t*   p) { return (intptr_t)specialized_load_acquire(p); }
+inline void*    OrderAccess::load_ptr_acquire(volatile void*       p) { return (void*)specialized_load_acquire((volatile intptr_t*)p); }
+inline void*    OrderAccess::load_ptr_acquire(const volatile void* p) { return (void*)specialized_load_acquire((volatile intptr_t*)p); }
+
+inline void     OrderAccess::release_store(volatile jbyte*   p, jbyte   v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jshort*  p, jshort  v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jint*    p, jint    v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jlong*   p, jlong   v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jfloat*  p, jfloat  v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jdouble* p, jdouble v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store(volatile jubyte*  p, jubyte  v) { specialized_release_store((volatile jbyte*) p, (jbyte) v); }
+inline void     OrderAccess::release_store(volatile jushort* p, jushort v) { specialized_release_store((volatile jshort*)p, (jshort)v); }
+inline void     OrderAccess::release_store(volatile juint*   p, juint   v) { specialized_release_store((volatile jint*)  p, (jint)  v); }
+inline void     OrderAccess::release_store(volatile julong*  p, julong  v) { specialized_release_store((volatile jlong*) p, (jlong) v); }
+
+inline void     OrderAccess::release_store_ptr(volatile intptr_t* p, intptr_t v) { specialized_release_store(p, v); }
+inline void     OrderAccess::release_store_ptr(volatile void*     p, void*    v) { specialized_release_store((volatile intptr_t*)p, (intptr_t)v); }
+
+inline void     OrderAccess::release_store_fence(volatile jbyte*   p, jbyte   v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jshort*  p, jshort  v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jint*    p, jint    v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jlong*   p, jlong   v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jfloat*  p, jfloat  v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jdouble* p, jdouble v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_fence(volatile jubyte*  p, jubyte  v) { specialized_release_store_fence((volatile jbyte*) p, (jbyte) v); }
+inline void     OrderAccess::release_store_fence(volatile jushort* p, jushort v) { specialized_release_store_fence((volatile jshort*)p, (jshort)v); }
+inline void     OrderAccess::release_store_fence(volatile juint*   p, juint   v) { specialized_release_store_fence((volatile jint*)  p, (jint)  v); }
+inline void     OrderAccess::release_store_fence(volatile julong*  p, julong  v) { specialized_release_store_fence((volatile jlong*) p, (jlong) v); }
+
+inline void     OrderAccess::release_store_ptr_fence(volatile intptr_t* p, intptr_t v) { specialized_release_store_fence(p, v); }
+inline void     OrderAccess::release_store_ptr_fence(volatile void*     p, void*    v) { specialized_release_store_fence((volatile intptr_t*)p, (intptr_t)v); }
+
+// The following methods can be specialized using simple template specialization
+// in the platform specific files for optimization purposes. Otherwise the
+// generalized variant is used.
+template<typename T> inline T    OrderAccess::specialized_load_acquire       (volatile T* p)       { return ordered_load<T, X_ACQUIRE>(p);    }
+template<typename T> inline void OrderAccess::specialized_release_store      (volatile T* p, T v)  { ordered_store<T, RELEASE_X>(p, v);       }
+template<typename T> inline void OrderAccess::specialized_release_store_fence(volatile T* p, T v)  { ordered_store<T, RELEASE_X_FENCE>(p, v); }
+
+// Generalized atomic volatile accesses valid in OrderAccess
+// All other types can be expressed in terms of these.
+inline void OrderAccess::store(volatile jbyte*   p, jbyte   v) { *p = v; }
+inline void OrderAccess::store(volatile jshort*  p, jshort  v) { *p = v; }
+inline void OrderAccess::store(volatile jint*    p, jint    v) { *p = v; }
+inline void OrderAccess::store(volatile jlong*   p, jlong   v) { Atomic::store(v, p); }
+inline void OrderAccess::store(volatile jdouble* p, jdouble v) { Atomic::store(jlong_cast(v), (volatile jlong*)p); }
+inline void OrderAccess::store(volatile jfloat*  p, jfloat  v) { *p = v; }
+
+inline jbyte   OrderAccess::load(volatile jbyte*   p) { return *p; }
+inline jshort  OrderAccess::load(volatile jshort*  p) { return *p; }
+inline jint    OrderAccess::load(volatile jint*    p) { return *p; }
+inline jlong   OrderAccess::load(volatile jlong*   p) { return Atomic::load(p); }
+inline jdouble OrderAccess::load(volatile jdouble* p) { return jdouble_cast(Atomic::load((volatile jlong*)p)); }
+inline jfloat  OrderAccess::load(volatile jfloat*  p) { return *p; }
+
+#endif // VM_HAS_GENERALIZED_ORDER_ACCESS
+
 #endif // SHARE_VM_RUNTIME_ORDERACCESS_INLINE_HPP
--- a/hotspot/src/share/vm/services/mallocSiteTable.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/services/mallocSiteTable.cpp	Thu Mar 12 16:10:58 2015 +0100
@@ -135,8 +135,7 @@
  */
 MallocSite* MallocSiteTable::lookup_or_add(const NativeCallStack& key, size_t* bucket_idx,
   size_t* pos_idx) {
-  int index = hash_to_index(key.hash());
-  assert(index >= 0, err_msg("Negative index %d", index));
+  unsigned int index = hash_to_index(key.hash());
   *bucket_idx = (size_t)index;
   *pos_idx = 0;
 
--- a/hotspot/src/share/vm/services/mallocSiteTable.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/services/mallocSiteTable.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -238,8 +238,7 @@
   static MallocSite* malloc_site(size_t bucket_idx, size_t pos_idx);
   static bool walk(MallocSiteWalker* walker);
 
-  static inline int hash_to_index(int  hash) {
-    hash = (hash > 0) ? hash : (-hash);
+  static inline unsigned int hash_to_index(unsigned int hash) {
     return (hash % table_size);
   }
 
--- a/hotspot/src/share/vm/utilities/globalDefinitions_sparcWorks.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/utilities/globalDefinitions_sparcWorks.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 1997, 2011, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 1997, 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
@@ -33,7 +33,9 @@
 
 
 # include <ctype.h>
+#define __USE_LEGACY_PROTOTYPES__
 # include <dirent.h>
+#undef __USE_LEGACY_PROTOTYPES__
 # include <string.h>
 # include <strings.h>     // for bsd'isms
 # include <stdarg.h>
--- a/hotspot/src/share/vm/utilities/nativeCallStack.cpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/utilities/nativeCallStack.cpp	Thu Mar 12 16:10:58 2015 +0100
@@ -55,6 +55,7 @@
   for (; index < NMT_TrackingStackDepth; index ++) {
     _stack[index] = NULL;
   }
+  _hash_value = 0;
 }
 
 // number of stack frames captured
@@ -69,19 +70,16 @@
 }
 
 // Hash code. Any better algorithm?
-int NativeCallStack::hash() const {
-  long hash_val = _hash_value;
+unsigned int NativeCallStack::hash() const {
+  uintptr_t hash_val = _hash_value;
   if (hash_val == 0) {
-    long pc;
-    int  index;
-    for (index = 0; index < NMT_TrackingStackDepth; index ++) {
-      pc = (long)_stack[index];
-      if (pc == 0) break;
-      hash_val += pc;
+    for (int index = 0; index < NMT_TrackingStackDepth; index++) {
+      if (_stack[index] == NULL) break;
+      hash_val += (uintptr_t)_stack[index];
     }
 
     NativeCallStack* p = const_cast<NativeCallStack*>(this);
-    p->_hash_value = (int)(hash_val & 0xFFFFFFFF);
+    p->_hash_value = (unsigned int)(hash_val & 0xFFFFFFFF);
   }
   return _hash_value;
 }
--- a/hotspot/src/share/vm/utilities/nativeCallStack.hpp	Mon Mar 09 09:30:16 2015 -0700
+++ b/hotspot/src/share/vm/utilities/nativeCallStack.hpp	Thu Mar 12 16:10:58 2015 +0100
@@ -56,8 +56,8 @@
   static const NativeCallStack EMPTY_STACK;
 
  private:
-  address   _stack[NMT_TrackingStackDepth];
-  int       _hash_value;
+  address       _stack[NMT_TrackingStackDepth];
+  unsigned int  _hash_value;
 
  public:
   NativeCallStack(int toSkip = 0, bool fillStack = false);
@@ -89,7 +89,7 @@
   }
 
   // Hash code. Any better algorithm?
-  int hash() const;
+  unsigned int hash() const;
 
   void print_on(outputStream* out) const;
   void print_on(outputStream* out, int indent) const;