--- a/.hgtags	Mon Nov 05 15:30:22 2012 -0500
+++ b/.hgtags	Mon Nov 05 13:55:31 2012 -0800
@@ -287,3 +287,7 @@
 b261523fe66c40a02968f0aa7e73602491bb3386 hs25-b05
 4547dc71db765276e027b0c2780b724bae0a07d3 jdk8-b61
 d0337c31c8be7716369b4e7c3bd5f352983c6a06 hs25-b06
+dccd40de8db1fa96f186e6179907818d75320440 jdk8-b62
+dc16fe422c535ecd4e9f80fb814a1bb9704da6f5 hs25-b07
+acabb5c282f59be7e3238920b2ea06b684ab68f7 jdk8-b63
+8cb93eadfb6dcab88d91b8e2cd3e0e07d0ac4048 hs25-b08

--- a/make/Makefile	Mon Nov 05 15:30:22 2012 -0500
+++ b/make/Makefile	Mon Nov 05 13:55:31 2012 -0800
@@ -453,14 +453,30 @@
     ifeq ($(JVM_VARIANT_ZEROSHARK), true)
         $(EXPORT_JRE_LIB_ARCH_DIR)/%.$(LIBRARY_SUFFIX): $(SHARK_DIR)/%.$(LIBRARY_SUFFIX)
 		$(install-file)
+        $(EXPORT_JRE_LIB_ARCH_DIR)/%.debuginfo):	$(SHARK_DIR)/%.debuginfo
+		$(install-file)
+        $(EXPORT_JRE_LIB_ARCH_DIR)/%.diz:		$(SHARK_DIR)/%.diz
+		$(install-file)
         $(EXPORT_SERVER_DIR)/%.$(LIBRARY_SUFFIX):       $(SHARK_DIR)/%.$(LIBRARY_SUFFIX)
 		$(install-file)
+        $(EXPORT_SERVER_DIR)/%.debuginfo:		$(SHARK_DIR)/%.debuginfo
+		$(install-file)
+        $(EXPORT_SERVER_DIR)/%.diz:			$(SHARK_DIR)/%.diz
+		$(install-file)
     endif
     ifeq ($(JVM_VARIANT_ZERO), true)
         $(EXPORT_JRE_LIB_ARCH_DIR)/%.$(LIBRARY_SUFFIX): $(ZERO_DIR)/%.$(LIBRARY_SUFFIX)
 		$(install-file)
+        $(EXPORT_JRE_LIB_ARCH_DIR)/%.debuginfo:		$(ZERO_DIR)/%.debuginfo
+		$(install-file)
+        $(EXPORT_JRE_LIB_ARCH_DIR)/%.diz:		$(ZERO_DIR)/%.diz
+		$(install-file)
         $(EXPORT_SERVER_DIR)/%.$(LIBRARY_SUFFIX):       $(ZERO_DIR)/%.$(LIBRARY_SUFFIX)
 		$(install-file)
+        $(EXPORT_SERVER_DIR)/%.debuginfo:		$(ZERO_DIR)/%.debuginfo
+		$(install-file)
+        $(EXPORT_SERVER_DIR)/%.diz:			$(ZERO_DIR)/%.diz
+		$(install-file)
     endif
     ifeq ($(JVM_VARIANT_MINIMAL1), true)
         $(EXPORT_JRE_LIB_ARCH_DIR)/%.$(LIBRARY_SUFFIX):	$(MINIMAL1_DIR)/%.$(LIBRARY_SUFFIX)

--- a/make/excludeSrc.make	Mon Nov 05 15:30:22 2012 -0500
+++ b/make/excludeSrc.make	Mon Nov 05 13:55:31 2012 -0800
@@ -79,10 +79,10 @@
       CXXFLAGS += -DSERIALGC
       CFLAGS += -DSERIALGC
       Src_Files_EXCLUDE += \
-	binaryTreeDictionary.cpp cmsAdaptiveSizePolicy.cpp cmsCollectorPolicy.cpp \
+	cmsAdaptiveSizePolicy.cpp cmsCollectorPolicy.cpp \
 	cmsGCAdaptivePolicyCounters.cpp cmsLockVerifier.cpp cmsPermGen.cpp compactibleFreeListSpace.cpp \
-	concurrentMarkSweepGeneration.cpp concurrentMarkSweepThread.cpp freeBlockDictionary.cpp \
-	freeChunk.cpp freeList.cpp promotionInfo.cpp vmCMSOperations.cpp collectionSetChooser.cpp \
+	concurrentMarkSweepGeneration.cpp concurrentMarkSweepThread.cpp \
+	freeChunk.cpp adaptiveFreeList.cpp promotionInfo.cpp vmCMSOperations.cpp collectionSetChooser.cpp \
 	concurrentG1Refine.cpp concurrentG1RefineThread.cpp concurrentMark.cpp concurrentMarkThread.cpp \
 	dirtyCardQueue.cpp g1AllocRegion.cpp g1BlockOffsetTable.cpp g1CollectedHeap.cpp g1GCPhaseTimes.cpp \
 	g1CollectorPolicy.cpp g1ErgoVerbose.cpp g1_globals.cpp g1HRPrinter.cpp g1MarkSweep.cpp \

--- a/make/hotspot_version	Mon Nov 05 15:30:22 2012 -0500
+++ b/make/hotspot_version	Mon Nov 05 13:55:31 2012 -0800
@@ -35,7 +35,7 @@
 
 HS_MAJOR_VER=25
 HS_MINOR_VER=0
-HS_BUILD_NUMBER=07
+HS_BUILD_NUMBER=09
 
 JDK_MAJOR_VER=1
 JDK_MINOR_VER=8

--- a/src/cpu/x86/vm/assembler_x86.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/assembler_x86.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -1007,6 +1007,67 @@
   emit_simd_arith(0x58, dst, src, VEX_SIMD_F3);
 }
 
+void Assembler::aesdec(XMMRegister dst, Address src) {
+  assert(VM_Version::supports_aes(), "");
+  InstructionMark im(this);
+  simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xde);
+  emit_operand(dst, src);
+}
+
+void Assembler::aesdec(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_aes(), "");
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xde);
+  emit_byte(0xC0 | encode);
+}
+
+void Assembler::aesdeclast(XMMRegister dst, Address src) {
+  assert(VM_Version::supports_aes(), "");
+  InstructionMark im(this);
+  simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdf);
+  emit_operand(dst, src);
+}
+
+void Assembler::aesdeclast(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_aes(), "");
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdf);
+  emit_byte(0xC0 | encode);
+}
+
+void Assembler::aesenc(XMMRegister dst, Address src) {
+  assert(VM_Version::supports_aes(), "");
+  InstructionMark im(this);
+  simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdc);
+  emit_operand(dst, src);
+}
+
+void Assembler::aesenc(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_aes(), "");
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdc);
+  emit_byte(0xC0 | encode);
+}
+
+void Assembler::aesenclast(XMMRegister dst, Address src) {
+  assert(VM_Version::supports_aes(), "");
+  InstructionMark im(this);
+  simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdd);
+  emit_operand(dst, src);
+}
+
+void Assembler::aesenclast(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_aes(), "");
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0xdd);
+  emit_byte(0xC0 | encode);
+}
+
+
 void Assembler::andl(Address dst, int32_t imm32) {
   InstructionMark im(this);
   prefix(dst);
@@ -2307,6 +2368,22 @@
   a_byte(p);
 }
 
+void Assembler::pshufb(XMMRegister dst, XMMRegister src) {
+  assert(VM_Version::supports_ssse3(), "");
+  int encode = simd_prefix_and_encode(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0x00);
+  emit_byte(0xC0 | encode);
+}
+
+void Assembler::pshufb(XMMRegister dst, Address src) {
+  assert(VM_Version::supports_ssse3(), "");
+  assert((UseAVX > 0), "SSE mode requires address alignment 16 bytes");
+  InstructionMark im(this);
+  simd_prefix(dst, dst, src, VEX_SIMD_66, VEX_OPCODE_0F_38);
+  emit_byte(0x00);
+  emit_operand(dst, src);
+}
+
 void Assembler::pshufd(XMMRegister dst, XMMRegister src, int mode) {
   assert(isByte(mode), "invalid value");
   NOT_LP64(assert(VM_Version::supports_sse2(), ""));
@@ -8067,6 +8144,15 @@
   LP64_ONLY(movq(dst, src)) NOT_LP64(movl(dst, src));
 }
 
+void MacroAssembler::movdqu(XMMRegister dst, AddressLiteral src) {
+  if (reachable(src)) {
+    Assembler::movdqu(dst, as_Address(src));
+  } else {
+    lea(rscratch1, src);
+    Assembler::movdqu(dst, Address(rscratch1, 0));
+  }
+}
+
 void MacroAssembler::movsd(XMMRegister dst, AddressLiteral src) {
   if (reachable(src)) {
     Assembler::movsd(dst, as_Address(src));
@@ -8357,6 +8443,17 @@
   }
 }
 
+void MacroAssembler::pshufb(XMMRegister dst, AddressLiteral src) {
+  // Used in sign-bit flipping with aligned address.
+  assert((UseAVX > 0) || (((intptr_t)src.target() & 15) == 0), "SSE mode requires address alignment 16 bytes");
+  if (reachable(src)) {
+    Assembler::pshufb(dst, as_Address(src));
+  } else {
+    lea(rscratch1, src);
+    Assembler::pshufb(dst, Address(rscratch1, 0));
+  }
+}
+
 // AVX 3-operands instructions
 
 void MacroAssembler::vaddsd(XMMRegister dst, XMMRegister nds, AddressLiteral src) {

--- a/src/cpu/x86/vm/assembler_x86.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/assembler_x86.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -875,6 +875,17 @@
   void addss(XMMRegister dst, Address src);
   void addss(XMMRegister dst, XMMRegister src);
 
+  // AES instructions
+  void aesdec(XMMRegister dst, Address src);
+  void aesdec(XMMRegister dst, XMMRegister src);
+  void aesdeclast(XMMRegister dst, Address src);
+  void aesdeclast(XMMRegister dst, XMMRegister src);
+  void aesenc(XMMRegister dst, Address src);
+  void aesenc(XMMRegister dst, XMMRegister src);
+  void aesenclast(XMMRegister dst, Address src);
+  void aesenclast(XMMRegister dst, XMMRegister src);
+
+
   void andl(Address  dst, int32_t imm32);
   void andl(Register dst, int32_t imm32);
   void andl(Register dst, Address src);
@@ -1424,6 +1435,10 @@
   void prefetcht2(Address src);
   void prefetchw(Address src);
 
+  // Shuffle Bytes
+  void pshufb(XMMRegister dst, XMMRegister src);
+  void pshufb(XMMRegister dst, Address src);
+
   // Shuffle Packed Doublewords
   void pshufd(XMMRegister dst, XMMRegister src, int mode);
   void pshufd(XMMRegister dst, Address src,     int mode);
@@ -2611,6 +2626,12 @@
   void divss(XMMRegister dst, Address src)        { Assembler::divss(dst, src); }
   void divss(XMMRegister dst, AddressLiteral src);
 
+  // Move Unaligned Double Quadword
+  void movdqu(Address     dst, XMMRegister src)   { Assembler::movdqu(dst, src); }
+  void movdqu(XMMRegister dst, Address src)       { Assembler::movdqu(dst, src); }
+  void movdqu(XMMRegister dst, XMMRegister src)   { Assembler::movdqu(dst, src); }
+  void movdqu(XMMRegister dst, AddressLiteral src);
+
   void movsd(XMMRegister dst, XMMRegister src) { Assembler::movsd(dst, src); }
   void movsd(Address dst, XMMRegister src)     { Assembler::movsd(dst, src); }
   void movsd(XMMRegister dst, Address src)     { Assembler::movsd(dst, src); }
@@ -2658,6 +2679,10 @@
   void xorps(XMMRegister dst, Address src)     { Assembler::xorps(dst, src); }
   void xorps(XMMRegister dst, AddressLiteral src);
 
+  // Shuffle Bytes
+  void pshufb(XMMRegister dst, XMMRegister src) { Assembler::pshufb(dst, src); }
+  void pshufb(XMMRegister dst, Address src)     { Assembler::pshufb(dst, src); }
+  void pshufb(XMMRegister dst, AddressLiteral src);
   // AVX 3-operands instructions
 
   void vaddsd(XMMRegister dst, XMMRegister nds, XMMRegister src) { Assembler::vaddsd(dst, nds, src); }

--- a/src/cpu/x86/vm/stubGenerator_x86_32.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubGenerator_x86_32.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -2137,6 +2137,529 @@
     }
   }
 
+  // AES intrinsic stubs
+  enum {AESBlockSize = 16};
+
+  address generate_key_shuffle_mask() {
+    __ align(16);
+    StubCodeMark mark(this, "StubRoutines", "key_shuffle_mask");
+    address start = __ pc();
+    __ emit_data(0x00010203, relocInfo::none, 0 );
+    __ emit_data(0x04050607, relocInfo::none, 0 );
+    __ emit_data(0x08090a0b, relocInfo::none, 0 );
+    __ emit_data(0x0c0d0e0f, relocInfo::none, 0 );
+    return start;
+  }
+
+  // Utility routine for loading a 128-bit key word in little endian format
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void load_key(XMMRegister xmmdst, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    __ movdqu(xmmdst, Address(key, offset));
+    if (xmm_shuf_mask != NULL) {
+      __ pshufb(xmmdst, xmm_shuf_mask);
+    } else {
+      __ pshufb(xmmdst, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    }
+  }
+
+  // aesenc using specified key+offset
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void aes_enc_key(XMMRegister xmmdst, XMMRegister xmmtmp, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    load_key(xmmtmp, key, offset, xmm_shuf_mask);
+    __ aesenc(xmmdst, xmmtmp);
+  }
+
+  // aesdec using specified key+offset
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void aes_dec_key(XMMRegister xmmdst, XMMRegister xmmtmp, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    load_key(xmmtmp, key, offset, xmm_shuf_mask);
+    __ aesdec(xmmdst, xmmtmp);
+  }
+
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //
+  address generate_aescrypt_encryptBlock() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "aescrypt_encryptBlock");
+    Label L_doLast;
+    address start = __ pc();
+
+    const Register from        = rsi;      // source array address
+    const Register to          = rdx;      // destination array address
+    const Register key         = rcx;      // key array address
+    const Register keylen      = rax;
+    const Address  from_param(rbp, 8+0);
+    const Address  to_param  (rbp, 8+4);
+    const Address  key_param (rbp, 8+8);
+
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    const XMMRegister xmm_key_shuf_mask = xmm2;
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+    __ push(rsi);
+    __ movptr(from , from_param);
+    __ movptr(to   , to_param);
+    __ movptr(key  , key_param);
+
+    __ movl(keylen, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    // keylen = # of 32-bit words, convert to 128-bit words
+    __ shrl(keylen, 2);
+    __ subl(keylen, 11);   // every key has at least 11 128-bit words, some have more
+
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    __ movdqu(xmm_result, Address(from, 0));  // get 16 bytes of input
+
+    // For encryption, the java expanded key ordering is just what we need
+
+    load_key(xmm_temp, key, 0x00, xmm_key_shuf_mask);
+    __ pxor(xmm_result, xmm_temp);
+    for (int offset = 0x10; offset <= 0x90; offset += 0x10) {
+      aes_enc_key(xmm_result, xmm_temp, key, offset, xmm_key_shuf_mask);
+    }
+    load_key  (xmm_temp, key, 0xa0, xmm_key_shuf_mask);
+    __ cmpl(keylen, 0);
+    __ jcc(Assembler::equal, L_doLast);
+    __ aesenc(xmm_result, xmm_temp);                   // only in 192 and 256 bit keys
+    aes_enc_key(xmm_result, xmm_temp, key, 0xb0, xmm_key_shuf_mask);
+    load_key(xmm_temp, key, 0xc0, xmm_key_shuf_mask);
+    __ subl(keylen, 2);
+    __ jcc(Assembler::equal, L_doLast);
+    __ aesenc(xmm_result, xmm_temp);                   // only in 256 bit keys
+    aes_enc_key(xmm_result, xmm_temp, key, 0xd0, xmm_key_shuf_mask);
+    load_key(xmm_temp, key, 0xe0, xmm_key_shuf_mask);
+
+    __ BIND(L_doLast);
+    __ aesenclast(xmm_result, xmm_temp);
+    __ movdqu(Address(to, 0), xmm_result);        // store the result
+    __ xorptr(rax, rax); // return 0
+    __ pop(rsi);
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+    return start;
+  }
+
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //
+  address generate_aescrypt_decryptBlock() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "aescrypt_decryptBlock");
+    Label L_doLast;
+    address start = __ pc();
+
+    const Register from        = rsi;      // source array address
+    const Register to          = rdx;      // destination array address
+    const Register key         = rcx;      // key array address
+    const Register keylen      = rax;
+    const Address  from_param(rbp, 8+0);
+    const Address  to_param  (rbp, 8+4);
+    const Address  key_param (rbp, 8+8);
+
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    const XMMRegister xmm_key_shuf_mask = xmm2;
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+    __ push(rsi);
+    __ movptr(from , from_param);
+    __ movptr(to   , to_param);
+    __ movptr(key  , key_param);
+
+    __ movl(keylen, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    // keylen = # of 32-bit words, convert to 128-bit words
+    __ shrl(keylen, 2);
+    __ subl(keylen, 11);   // every key has at least 11 128-bit words, some have more
+
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    __ movdqu(xmm_result, Address(from, 0));
+
+    // for decryption java expanded key ordering is rotated one position from what we want
+    // so we start from 0x10 here and hit 0x00 last
+    // we don't know if the key is aligned, hence not using load-execute form
+    load_key(xmm_temp, key, 0x10, xmm_key_shuf_mask);
+    __ pxor  (xmm_result, xmm_temp);
+    for (int offset = 0x20; offset <= 0xa0; offset += 0x10) {
+      aes_dec_key(xmm_result, xmm_temp, key, offset, xmm_key_shuf_mask);
+    }
+    __ cmpl(keylen, 0);
+    __ jcc(Assembler::equal, L_doLast);
+    // only in 192 and 256 bit keys
+    aes_dec_key(xmm_result, xmm_temp, key, 0xb0, xmm_key_shuf_mask);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xc0, xmm_key_shuf_mask);
+    __ subl(keylen, 2);
+    __ jcc(Assembler::equal, L_doLast);
+    // only in 256 bit keys
+    aes_dec_key(xmm_result, xmm_temp, key, 0xd0, xmm_key_shuf_mask);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xe0, xmm_key_shuf_mask);
+
+    __ BIND(L_doLast);
+    // for decryption the aesdeclast operation is always on key+0x00
+    load_key(xmm_temp, key, 0x00, xmm_key_shuf_mask);
+    __ aesdeclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, 0), xmm_result);  // store the result
+
+    __ xorptr(rax, rax); // return 0
+    __ pop(rsi);
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+    return start;
+  }
+
+  void handleSOERegisters(bool saving) {
+    const int saveFrameSizeInBytes = 4 * wordSize;
+    const Address saved_rbx     (rbp, -3 * wordSize);
+    const Address saved_rsi     (rbp, -2 * wordSize);
+    const Address saved_rdi     (rbp, -1 * wordSize);
+
+    if (saving) {
+      __ subptr(rsp, saveFrameSizeInBytes);
+      __ movptr(saved_rsi, rsi);
+      __ movptr(saved_rdi, rdi);
+      __ movptr(saved_rbx, rbx);
+    } else {
+      // restoring
+      __ movptr(rsi, saved_rsi);
+      __ movptr(rdi, saved_rdi);
+      __ movptr(rbx, saved_rbx);
+    }
+  }
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //   c_rarg3   - r vector byte array address
+  //   c_rarg4   - input length
+  //
+  address generate_cipherBlockChaining_encryptAESCrypt() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "cipherBlockChaining_encryptAESCrypt");
+    address start = __ pc();
+
+    Label L_exit, L_key_192_256, L_key_256, L_loopTop_128, L_loopTop_192, L_loopTop_256;
+    const Register from        = rsi;      // source array address
+    const Register to          = rdx;      // destination array address
+    const Register key         = rcx;      // key array address
+    const Register rvec        = rdi;      // r byte array initialized from initvector array address
+                                           // and left with the results of the last encryption block
+    const Register len_reg     = rbx;      // src len (must be multiple of blocksize 16)
+    const Register pos         = rax;
+
+    // xmm register assignments for the loops below
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    // first 6 keys preloaded into xmm2-xmm7
+    const int XMM_REG_NUM_KEY_FIRST = 2;
+    const int XMM_REG_NUM_KEY_LAST  = 7;
+    const XMMRegister xmm_key0   = as_XMMRegister(XMM_REG_NUM_KEY_FIRST);
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+    handleSOERegisters(true /*saving*/);
+
+    // load registers from incoming parameters
+    const Address  from_param(rbp, 8+0);
+    const Address  to_param  (rbp, 8+4);
+    const Address  key_param (rbp, 8+8);
+    const Address  rvec_param (rbp, 8+12);
+    const Address  len_param  (rbp, 8+16);
+    __ movptr(from , from_param);
+    __ movptr(to   , to_param);
+    __ movptr(key  , key_param);
+    __ movptr(rvec , rvec_param);
+    __ movptr(len_reg , len_param);
+
+    const XMMRegister xmm_key_shuf_mask = xmm_temp;  // used temporarily to swap key bytes up front
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    // load up xmm regs 2 thru 7 with keys 0-5
+    for (int rnum = XMM_REG_NUM_KEY_FIRST, offset = 0x00; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      load_key(as_XMMRegister(rnum), key, offset, xmm_key_shuf_mask);
+      offset += 0x10;
+    }
+
+    __ movdqu(xmm_result, Address(rvec, 0x00));   // initialize xmm_result with r vec
+
+    // now split to different paths depending on the keylen (len in ints of AESCrypt.KLE array (52=192, or 60=256))
+    __ movl(rax, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    __ cmpl(rax, 44);
+    __ jcc(Assembler::notEqual, L_key_192_256);
+
+    // 128 bit code follows here
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_loopTop_128);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);                                // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);                                // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = 0x60; key_offset <= 0x90; key_offset += 0x10) {
+      aes_enc_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0xa0);
+    __ aesenclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_128);
+
+    __ BIND(L_exit);
+    __ movdqu(Address(rvec, 0), xmm_result);     // final value of r stored in rvec of CipherBlockChaining object
+
+    handleSOERegisters(false /*restoring*/);
+    __ movl(rax, 0);                             // return 0 (why?)
+    __ leave();                                  // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+  __ BIND(L_key_192_256);
+  // here rax = len in ints of AESCrypt.KLE array (52=192, or 60=256)
+    __ cmpl(rax, 52);
+    __ jcc(Assembler::notEqual, L_key_256);
+
+    // 192-bit code follows here (could be changed to use more xmm registers)
+    __ movptr(pos, 0);
+  __ align(OptoLoopAlignment);
+  __ BIND(L_loopTop_192);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);                                // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);                                // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = 0x60; key_offset <= 0xb0; key_offset += 0x10) {
+      aes_enc_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0xc0);
+    __ aesenclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);   // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_192);
+    __ jmp(L_exit);
+
+  __ BIND(L_key_256);
+    // 256-bit code follows here (could be changed to use more xmm registers)
+    __ movptr(pos, 0);
+  __ align(OptoLoopAlignment);
+  __ BIND(L_loopTop_256);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);                                // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);                                // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = 0x60; key_offset <= 0xd0; key_offset += 0x10) {
+      aes_enc_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0xe0);
+    __ aesenclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);   // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_256);
+    __ jmp(L_exit);
+
+    return start;
+  }
+
+
+  // CBC AES Decryption.
+  // In 32-bit stub, because of lack of registers we do not try to parallelize 4 blocks at a time.
+  //
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //   c_rarg3   - r vector byte array address
+  //   c_rarg4   - input length
+  //
+
+  address generate_cipherBlockChaining_decryptAESCrypt() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "cipherBlockChaining_decryptAESCrypt");
+    address start = __ pc();
+
+    Label L_exit, L_key_192_256, L_key_256;
+    Label L_singleBlock_loopTop_128;
+    Label L_singleBlock_loopTop_192, L_singleBlock_loopTop_256;
+    const Register from        = rsi;      // source array address
+    const Register to          = rdx;      // destination array address
+    const Register key         = rcx;      // key array address
+    const Register rvec        = rdi;      // r byte array initialized from initvector array address
+                                           // and left with the results of the last encryption block
+    const Register len_reg     = rbx;      // src len (must be multiple of blocksize 16)
+    const Register pos         = rax;
+
+    // xmm register assignments for the loops below
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    // first 6 keys preloaded into xmm2-xmm7
+    const int XMM_REG_NUM_KEY_FIRST = 2;
+    const int XMM_REG_NUM_KEY_LAST  = 7;
+    const int FIRST_NON_REG_KEY_offset = 0x70;
+    const XMMRegister xmm_key_first   = as_XMMRegister(XMM_REG_NUM_KEY_FIRST);
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+    handleSOERegisters(true /*saving*/);
+
+    // load registers from incoming parameters
+    const Address  from_param(rbp, 8+0);
+    const Address  to_param  (rbp, 8+4);
+    const Address  key_param (rbp, 8+8);
+    const Address  rvec_param (rbp, 8+12);
+    const Address  len_param  (rbp, 8+16);
+    __ movptr(from , from_param);
+    __ movptr(to   , to_param);
+    __ movptr(key  , key_param);
+    __ movptr(rvec , rvec_param);
+    __ movptr(len_reg , len_param);
+
+    // the java expanded key ordering is rotated one position from what we want
+    // so we start from 0x10 here and hit 0x00 last
+    const XMMRegister xmm_key_shuf_mask = xmm1;  // used temporarily to swap key bytes up front
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    // load up xmm regs 2 thru 6 with first 5 keys
+    for (int rnum = XMM_REG_NUM_KEY_FIRST, offset = 0x10; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      load_key(as_XMMRegister(rnum), key, offset, xmm_key_shuf_mask);
+      offset += 0x10;
+    }
+
+    // inside here, use the rvec register to point to previous block cipher
+    // with which we xor at the end of each newly decrypted block
+    const Register  prev_block_cipher_ptr = rvec;
+
+    // now split to different paths depending on the keylen (len in ints of AESCrypt.KLE array (52=192, or 60=256))
+    __ movl(rax, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    __ cmpl(rax, 44);
+    __ jcc(Assembler::notEqual, L_key_192_256);
+
+
+    // 128-bit code follows here, parallelized
+    __ movptr(pos, 0);
+  __ align(OptoLoopAlignment);
+  __ BIND(L_singleBlock_loopTop_128);
+    __ cmpptr(len_reg, 0);           // any blocks left??
+    __ jcc(Assembler::equal, L_exit);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ pxor  (xmm_result, xmm_key_first);                             // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = FIRST_NON_REG_KEY_offset; key_offset <= 0xa0; key_offset += 0x10) {   // 128-bit runs up to key offset a0
+      aes_dec_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0x00);                                     // final key is stored in java expanded array at offset 0
+    __ aesdeclast(xmm_result, xmm_temp);
+    __ movdqu(xmm_temp, Address(prev_block_cipher_ptr, 0x00));
+    __ pxor  (xmm_result, xmm_temp);                                  // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ lea(prev_block_cipher_ptr, Address(from, pos, Address::times_1, 0));     // set up new ptr
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jmp(L_singleBlock_loopTop_128);
+
+
+    __ BIND(L_exit);
+    __ movdqu(xmm_temp, Address(prev_block_cipher_ptr, 0x00));
+    __ movptr(rvec , rvec_param);                                     // restore this since used in loop
+    __ movdqu(Address(rvec, 0), xmm_temp);                            // final value of r stored in rvec of CipherBlockChaining object
+    handleSOERegisters(false /*restoring*/);
+    __ movl(rax, 0);                                                  // return 0 (why?)
+    __ leave();                                                       // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+
+    __ BIND(L_key_192_256);
+    // here rax = len in ints of AESCrypt.KLE array (52=192, or 60=256)
+    __ cmpl(rax, 52);
+    __ jcc(Assembler::notEqual, L_key_256);
+
+    // 192-bit code follows here (could be optimized to use parallelism)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_singleBlock_loopTop_192);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ pxor  (xmm_result, xmm_key_first);                             // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = FIRST_NON_REG_KEY_offset; key_offset <= 0xc0; key_offset += 0x10) {   // 192-bit runs up to key offset c0
+      aes_dec_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0x00);                                     // final key is stored in java expanded array at offset 0
+    __ aesdeclast(xmm_result, xmm_temp);
+    __ movdqu(xmm_temp, Address(prev_block_cipher_ptr, 0x00));
+    __ pxor  (xmm_result, xmm_temp);                                  // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ lea(prev_block_cipher_ptr, Address(from, pos, Address::times_1, 0));     // set up new ptr
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual,L_singleBlock_loopTop_192);
+    __ jmp(L_exit);
+
+    __ BIND(L_key_256);
+    // 256-bit code follows here (could be optimized to use parallelism)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_singleBlock_loopTop_256);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ pxor  (xmm_result, xmm_key_first);                             // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    for (int key_offset = FIRST_NON_REG_KEY_offset; key_offset <= 0xe0; key_offset += 0x10) {   // 256-bit runs up to key offset e0
+      aes_dec_key(xmm_result, xmm_temp, key, key_offset);
+    }
+    load_key(xmm_temp, key, 0x00);                                     // final key is stored in java expanded array at offset 0
+    __ aesdeclast(xmm_result, xmm_temp);
+    __ movdqu(xmm_temp, Address(prev_block_cipher_ptr, 0x00));
+    __ pxor  (xmm_result, xmm_temp);                                  // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ lea(prev_block_cipher_ptr, Address(from, pos, Address::times_1, 0));     // set up new ptr
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual,L_singleBlock_loopTop_256);
+    __ jmp(L_exit);
+
+    return start;
+  }
+
+
  public:
   // Information about frame layout at time of blocking runtime call.
   // Note that we only have to preserve callee-saved registers since
@@ -2332,6 +2855,16 @@
     generate_arraycopy_stubs();
 
     generate_math_stubs();
+
+    // don't bother generating these AES intrinsic stubs unless global flag is set
+    if (UseAESIntrinsics) {
+      StubRoutines::x86::_key_shuffle_mask_addr = generate_key_shuffle_mask();  // might be needed by the others
+
+      StubRoutines::_aescrypt_encryptBlock = generate_aescrypt_encryptBlock();
+      StubRoutines::_aescrypt_decryptBlock = generate_aescrypt_decryptBlock();
+      StubRoutines::_cipherBlockChaining_encryptAESCrypt = generate_cipherBlockChaining_encryptAESCrypt();
+      StubRoutines::_cipherBlockChaining_decryptAESCrypt = generate_cipherBlockChaining_decryptAESCrypt();
+    }
   }
 
 

--- a/src/cpu/x86/vm/stubGenerator_x86_64.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubGenerator_x86_64.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -2941,6 +2941,548 @@
     }
   }
 
+  // AES intrinsic stubs
+  enum {AESBlockSize = 16};
+
+  address generate_key_shuffle_mask() {
+    __ align(16);
+    StubCodeMark mark(this, "StubRoutines", "key_shuffle_mask");
+    address start = __ pc();
+    __ emit_data64( 0x0405060700010203, relocInfo::none );
+    __ emit_data64( 0x0c0d0e0f08090a0b, relocInfo::none );
+    return start;
+  }
+
+  // Utility routine for loading a 128-bit key word in little endian format
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void load_key(XMMRegister xmmdst, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    __ movdqu(xmmdst, Address(key, offset));
+    if (xmm_shuf_mask != NULL) {
+      __ pshufb(xmmdst, xmm_shuf_mask);
+    } else {
+      __ pshufb(xmmdst, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    }
+  }
+
+  // aesenc using specified key+offset
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void aes_enc_key(XMMRegister xmmdst, XMMRegister xmmtmp, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    load_key(xmmtmp, key, offset, xmm_shuf_mask);
+    __ aesenc(xmmdst, xmmtmp);
+  }
+
+  // aesdec using specified key+offset
+  // can optionally specify that the shuffle mask is already in an xmmregister
+  void aes_dec_key(XMMRegister xmmdst, XMMRegister xmmtmp, Register key, int offset, XMMRegister xmm_shuf_mask=NULL) {
+    load_key(xmmtmp, key, offset, xmm_shuf_mask);
+    __ aesdec(xmmdst, xmmtmp);
+  }
+
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //
+  address generate_aescrypt_encryptBlock() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "aescrypt_encryptBlock");
+    Label L_doLast;
+    address start = __ pc();
+
+    const Register from        = c_rarg0;  // source array address
+    const Register to          = c_rarg1;  // destination array address
+    const Register key         = c_rarg2;  // key array address
+    const Register keylen      = rax;
+
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    const XMMRegister xmm_key_shuf_mask = xmm2;
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+
+    __ movl(keylen, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    // keylen = # of 32-bit words, convert to 128-bit words
+    __ shrl(keylen, 2);
+    __ subl(keylen, 11);   // every key has at least 11 128-bit words, some have more
+
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    __ movdqu(xmm_result, Address(from, 0));  // get 16 bytes of input
+
+    // For encryption, the java expanded key ordering is just what we need
+    // we don't know if the key is aligned, hence not using load-execute form
+
+    load_key(xmm_temp, key, 0x00, xmm_key_shuf_mask);
+    __ pxor(xmm_result, xmm_temp);
+    for (int offset = 0x10; offset <= 0x90; offset += 0x10) {
+      aes_enc_key(xmm_result, xmm_temp, key, offset, xmm_key_shuf_mask);
+    }
+    load_key  (xmm_temp, key, 0xa0, xmm_key_shuf_mask);
+    __ cmpl(keylen, 0);
+    __ jcc(Assembler::equal, L_doLast);
+    __ aesenc(xmm_result, xmm_temp);                   // only in 192 and 256 bit keys
+    aes_enc_key(xmm_result, xmm_temp, key, 0xb0, xmm_key_shuf_mask);
+    load_key(xmm_temp, key, 0xc0, xmm_key_shuf_mask);
+    __ subl(keylen, 2);
+    __ jcc(Assembler::equal, L_doLast);
+    __ aesenc(xmm_result, xmm_temp);                   // only in 256 bit keys
+    aes_enc_key(xmm_result, xmm_temp, key, 0xd0, xmm_key_shuf_mask);
+    load_key(xmm_temp, key, 0xe0, xmm_key_shuf_mask);
+
+    __ BIND(L_doLast);
+    __ aesenclast(xmm_result, xmm_temp);
+    __ movdqu(Address(to, 0), xmm_result);        // store the result
+    __ xorptr(rax, rax); // return 0
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+    return start;
+  }
+
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //
+  address generate_aescrypt_decryptBlock() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "aescrypt_decryptBlock");
+    Label L_doLast;
+    address start = __ pc();
+
+    const Register from        = c_rarg0;  // source array address
+    const Register to          = c_rarg1;  // destination array address
+    const Register key         = c_rarg2;  // key array address
+    const Register keylen      = rax;
+
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    const XMMRegister xmm_key_shuf_mask = xmm2;
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+
+    __ movl(keylen, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    // keylen = # of 32-bit words, convert to 128-bit words
+    __ shrl(keylen, 2);
+    __ subl(keylen, 11);   // every key has at least 11 128-bit words, some have more
+
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    __ movdqu(xmm_result, Address(from, 0));
+
+    // for decryption java expanded key ordering is rotated one position from what we want
+    // so we start from 0x10 here and hit 0x00 last
+    // we don't know if the key is aligned, hence not using load-execute form
+    load_key(xmm_temp, key, 0x10, xmm_key_shuf_mask);
+    __ pxor  (xmm_result, xmm_temp);
+    for (int offset = 0x20; offset <= 0xa0; offset += 0x10) {
+      aes_dec_key(xmm_result, xmm_temp, key, offset, xmm_key_shuf_mask);
+    }
+    __ cmpl(keylen, 0);
+    __ jcc(Assembler::equal, L_doLast);
+    // only in 192 and 256 bit keys
+    aes_dec_key(xmm_result, xmm_temp, key, 0xb0, xmm_key_shuf_mask);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xc0, xmm_key_shuf_mask);
+    __ subl(keylen, 2);
+    __ jcc(Assembler::equal, L_doLast);
+    // only in 256 bit keys
+    aes_dec_key(xmm_result, xmm_temp, key, 0xd0, xmm_key_shuf_mask);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xe0, xmm_key_shuf_mask);
+
+    __ BIND(L_doLast);
+    // for decryption the aesdeclast operation is always on key+0x00
+    load_key(xmm_temp, key, 0x00, xmm_key_shuf_mask);
+    __ aesdeclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, 0), xmm_result);  // store the result
+
+    __ xorptr(rax, rax); // return 0
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+    return start;
+  }
+
+
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //   c_rarg3   - r vector byte array address
+  //   c_rarg4   - input length
+  //
+  address generate_cipherBlockChaining_encryptAESCrypt() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "cipherBlockChaining_encryptAESCrypt");
+    address start = __ pc();
+
+    Label L_exit, L_key_192_256, L_key_256, L_loopTop_128, L_loopTop_192, L_loopTop_256;
+    const Register from        = c_rarg0;  // source array address
+    const Register to          = c_rarg1;  // destination array address
+    const Register key         = c_rarg2;  // key array address
+    const Register rvec        = c_rarg3;  // r byte array initialized from initvector array address
+                                           // and left with the results of the last encryption block
+#ifndef _WIN64
+    const Register len_reg     = c_rarg4;  // src len (must be multiple of blocksize 16)
+#else
+    const Address  len_mem(rsp, 6 * wordSize);  // length is on stack on Win64
+    const Register len_reg     = r10;      // pick the first volatile windows register
+#endif
+    const Register pos         = rax;
+
+    // xmm register assignments for the loops below
+    const XMMRegister xmm_result = xmm0;
+    const XMMRegister xmm_temp   = xmm1;
+    // keys 0-10 preloaded into xmm2-xmm12
+    const int XMM_REG_NUM_KEY_FIRST = 2;
+    const int XMM_REG_NUM_KEY_LAST  = 12;
+    const XMMRegister xmm_key0   = as_XMMRegister(XMM_REG_NUM_KEY_FIRST);
+    const XMMRegister xmm_key10  = as_XMMRegister(XMM_REG_NUM_KEY_LAST);
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+
+#ifdef _WIN64
+    // on win64, fill len_reg from stack position
+    __ movl(len_reg, len_mem);
+    // save the xmm registers which must be preserved 6-12
+    __ subptr(rsp, -rsp_after_call_off * wordSize);
+    for (int i = 6; i <= XMM_REG_NUM_KEY_LAST; i++) {
+      __ movdqu(xmm_save(i), as_XMMRegister(i));
+    }
+#endif
+
+    const XMMRegister xmm_key_shuf_mask = xmm_temp;  // used temporarily to swap key bytes up front
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    // load up xmm regs 2 thru 12 with key 0x00 - 0xa0
+    for (int rnum = XMM_REG_NUM_KEY_FIRST, offset = 0x00; rnum <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      load_key(as_XMMRegister(rnum), key, offset, xmm_key_shuf_mask);
+      offset += 0x10;
+    }
+
+    __ movdqu(xmm_result, Address(rvec, 0x00));   // initialize xmm_result with r vec
+
+    // now split to different paths depending on the keylen (len in ints of AESCrypt.KLE array (52=192, or 60=256))
+    __ movl(rax, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    __ cmpl(rax, 44);
+    __ jcc(Assembler::notEqual, L_key_192_256);
+
+    // 128 bit code follows here
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_loopTop_128);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);               // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);               // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum <= XMM_REG_NUM_KEY_LAST - 1; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    __ aesenclast(xmm_result, xmm_key10);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_128);
+
+    __ BIND(L_exit);
+    __ movdqu(Address(rvec, 0), xmm_result);     // final value of r stored in rvec of CipherBlockChaining object
+
+#ifdef _WIN64
+    // restore xmm regs belonging to calling function
+    for (int i = 6; i <= XMM_REG_NUM_KEY_LAST; i++) {
+      __ movdqu(as_XMMRegister(i), xmm_save(i));
+    }
+#endif
+    __ movl(rax, 0); // return 0 (why?)
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+    __ BIND(L_key_192_256);
+    // here rax = len in ints of AESCrypt.KLE array (52=192, or 60=256)
+    __ cmpl(rax, 52);
+    __ jcc(Assembler::notEqual, L_key_256);
+
+    // 192-bit code follows here (could be changed to use more xmm registers)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_loopTop_192);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);               // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);               // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    aes_enc_key(xmm_result, xmm_temp, key, 0xb0);
+    load_key(xmm_temp, key, 0xc0);
+    __ aesenclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_192);
+    __ jmp(L_exit);
+
+    __ BIND(L_key_256);
+    // 256-bit code follows here (could be changed to use more xmm registers)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_loopTop_256);
+    __ movdqu(xmm_temp, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of input
+    __ pxor  (xmm_result, xmm_temp);               // xor with the current r vector
+
+    __ pxor  (xmm_result, xmm_key0);               // do the aes rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      __ aesenc(xmm_result, as_XMMRegister(rnum));
+    }
+    aes_enc_key(xmm_result, xmm_temp, key, 0xb0);
+    aes_enc_key(xmm_result, xmm_temp, key, 0xc0);
+    aes_enc_key(xmm_result, xmm_temp, key, 0xd0);
+    load_key(xmm_temp, key, 0xe0);
+    __ aesenclast(xmm_result, xmm_temp);
+
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual, L_loopTop_256);
+    __ jmp(L_exit);
+
+    return start;
+  }
+
+
+
+  // This is a version of CBC/AES Decrypt which does 4 blocks in a loop at a time
+  // to hide instruction latency
+  //
+  // Arguments:
+  //
+  // Inputs:
+  //   c_rarg0   - source byte array address
+  //   c_rarg1   - destination byte array address
+  //   c_rarg2   - K (key) in little endian int array
+  //   c_rarg3   - r vector byte array address
+  //   c_rarg4   - input length
+  //
+
+  address generate_cipherBlockChaining_decryptAESCrypt_Parallel() {
+    assert(UseAES && (UseAVX > 0), "need AES instructions and misaligned SSE support");
+    __ align(CodeEntryAlignment);
+    StubCodeMark mark(this, "StubRoutines", "cipherBlockChaining_decryptAESCrypt");
+    address start = __ pc();
+
+    Label L_exit, L_key_192_256, L_key_256;
+    Label L_singleBlock_loopTop_128, L_multiBlock_loopTop_128;
+    Label L_singleBlock_loopTop_192, L_singleBlock_loopTop_256;
+    const Register from        = c_rarg0;  // source array address
+    const Register to          = c_rarg1;  // destination array address
+    const Register key         = c_rarg2;  // key array address
+    const Register rvec        = c_rarg3;  // r byte array initialized from initvector array address
+                                           // and left with the results of the last encryption block
+#ifndef _WIN64
+    const Register len_reg     = c_rarg4;  // src len (must be multiple of blocksize 16)
+#else
+    const Address  len_mem(rsp, 6 * wordSize);  // length is on stack on Win64
+    const Register len_reg     = r10;      // pick the first volatile windows register
+#endif
+    const Register pos         = rax;
+
+    // xmm register assignments for the loops below
+    const XMMRegister xmm_result = xmm0;
+    // keys 0-10 preloaded into xmm2-xmm12
+    const int XMM_REG_NUM_KEY_FIRST = 5;
+    const int XMM_REG_NUM_KEY_LAST  = 15;
+    const XMMRegister xmm_key_first   = as_XMMRegister(XMM_REG_NUM_KEY_FIRST);
+    const XMMRegister xmm_key_last  = as_XMMRegister(XMM_REG_NUM_KEY_LAST);
+
+    __ enter(); // required for proper stackwalking of RuntimeStub frame
+
+#ifdef _WIN64
+    // on win64, fill len_reg from stack position
+    __ movl(len_reg, len_mem);
+    // save the xmm registers which must be preserved 6-15
+    __ subptr(rsp, -rsp_after_call_off * wordSize);
+    for (int i = 6; i <= XMM_REG_NUM_KEY_LAST; i++) {
+      __ movdqu(xmm_save(i), as_XMMRegister(i));
+    }
+#endif
+    // the java expanded key ordering is rotated one position from what we want
+    // so we start from 0x10 here and hit 0x00 last
+    const XMMRegister xmm_key_shuf_mask = xmm1;  // used temporarily to swap key bytes up front
+    __ movdqu(xmm_key_shuf_mask, ExternalAddress(StubRoutines::x86::key_shuffle_mask_addr()));
+    // load up xmm regs 5 thru 15 with key 0x10 - 0xa0 - 0x00
+    for (int rnum = XMM_REG_NUM_KEY_FIRST, offset = 0x10; rnum <= XMM_REG_NUM_KEY_LAST; rnum++) {
+      if (rnum == XMM_REG_NUM_KEY_LAST) offset = 0x00;
+      load_key(as_XMMRegister(rnum), key, offset, xmm_key_shuf_mask);
+      offset += 0x10;
+    }
+
+    const XMMRegister xmm_prev_block_cipher = xmm1;  // holds cipher of previous block
+    // registers holding the four results in the parallelized loop
+    const XMMRegister xmm_result0 = xmm0;
+    const XMMRegister xmm_result1 = xmm2;
+    const XMMRegister xmm_result2 = xmm3;
+    const XMMRegister xmm_result3 = xmm4;
+
+    __ movdqu(xmm_prev_block_cipher, Address(rvec, 0x00));   // initialize with initial rvec
+
+    // now split to different paths depending on the keylen (len in ints of AESCrypt.KLE array (52=192, or 60=256))
+    __ movl(rax, Address(key, arrayOopDesc::length_offset_in_bytes() - arrayOopDesc::base_offset_in_bytes(T_INT)));
+    __ cmpl(rax, 44);
+    __ jcc(Assembler::notEqual, L_key_192_256);
+
+
+    // 128-bit code follows here, parallelized
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_multiBlock_loopTop_128);
+    __ cmpptr(len_reg, 4*AESBlockSize);           // see if at least 4 blocks left
+    __ jcc(Assembler::less, L_singleBlock_loopTop_128);
+
+    __ movdqu(xmm_result0, Address(from, pos, Address::times_1, 0*AESBlockSize));   // get next 4 blocks into xmmresult registers
+    __ movdqu(xmm_result1, Address(from, pos, Address::times_1, 1*AESBlockSize));
+    __ movdqu(xmm_result2, Address(from, pos, Address::times_1, 2*AESBlockSize));
+    __ movdqu(xmm_result3, Address(from, pos, Address::times_1, 3*AESBlockSize));
+
+#define DoFour(opc, src_reg)                    \
+    __ opc(xmm_result0, src_reg);               \
+    __ opc(xmm_result1, src_reg);               \
+    __ opc(xmm_result2, src_reg);               \
+    __ opc(xmm_result3, src_reg);
+
+    DoFour(pxor, xmm_key_first);
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST - 1; rnum++) {
+      DoFour(aesdec, as_XMMRegister(rnum));
+    }
+    DoFour(aesdeclast, xmm_key_last);
+    // for each result, xor with the r vector of previous cipher block
+    __ pxor(xmm_result0, xmm_prev_block_cipher);
+    __ movdqu(xmm_prev_block_cipher, Address(from, pos, Address::times_1, 0*AESBlockSize));
+    __ pxor(xmm_result1, xmm_prev_block_cipher);
+    __ movdqu(xmm_prev_block_cipher, Address(from, pos, Address::times_1, 1*AESBlockSize));
+    __ pxor(xmm_result2, xmm_prev_block_cipher);
+    __ movdqu(xmm_prev_block_cipher, Address(from, pos, Address::times_1, 2*AESBlockSize));
+    __ pxor(xmm_result3, xmm_prev_block_cipher);
+    __ movdqu(xmm_prev_block_cipher, Address(from, pos, Address::times_1, 3*AESBlockSize));   // this will carry over to next set of blocks
+
+    __ movdqu(Address(to, pos, Address::times_1, 0*AESBlockSize), xmm_result0);     // store 4 results into the next 64 bytes of output
+    __ movdqu(Address(to, pos, Address::times_1, 1*AESBlockSize), xmm_result1);
+    __ movdqu(Address(to, pos, Address::times_1, 2*AESBlockSize), xmm_result2);
+    __ movdqu(Address(to, pos, Address::times_1, 3*AESBlockSize), xmm_result3);
+
+    __ addptr(pos, 4*AESBlockSize);
+    __ subptr(len_reg, 4*AESBlockSize);
+    __ jmp(L_multiBlock_loopTop_128);
+
+    // registers used in the non-parallelized loops
+    const XMMRegister xmm_prev_block_cipher_save = xmm2;
+    const XMMRegister xmm_temp   = xmm3;
+
+    __ align(OptoLoopAlignment);
+    __ BIND(L_singleBlock_loopTop_128);
+    __ cmpptr(len_reg, 0);           // any blocks left??
+    __ jcc(Assembler::equal, L_exit);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ movdqa(xmm_prev_block_cipher_save, xmm_result);              // save for next r vector
+    __ pxor  (xmm_result, xmm_key_first);               // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum  <= XMM_REG_NUM_KEY_LAST - 1; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    __ aesdeclast(xmm_result, xmm_key_last);
+    __ pxor  (xmm_result, xmm_prev_block_cipher);               // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ movdqa(xmm_prev_block_cipher, xmm_prev_block_cipher_save);              // set up next r vector with cipher input from this block
+
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jmp(L_singleBlock_loopTop_128);
+
+
+    __ BIND(L_exit);
+    __ movdqu(Address(rvec, 0), xmm_prev_block_cipher);     // final value of r stored in rvec of CipherBlockChaining object
+#ifdef _WIN64
+    // restore regs belonging to calling function
+    for (int i = 6; i <= XMM_REG_NUM_KEY_LAST; i++) {
+      __ movdqu(as_XMMRegister(i), xmm_save(i));
+    }
+#endif
+    __ movl(rax, 0); // return 0 (why?)
+    __ leave(); // required for proper stackwalking of RuntimeStub frame
+    __ ret(0);
+
+
+    __ BIND(L_key_192_256);
+    // here rax = len in ints of AESCrypt.KLE array (52=192, or 60=256)
+    __ cmpl(rax, 52);
+    __ jcc(Assembler::notEqual, L_key_256);
+
+    // 192-bit code follows here (could be optimized to use parallelism)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_singleBlock_loopTop_192);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ movdqa(xmm_prev_block_cipher_save, xmm_result);              // save for next r vector
+    __ pxor  (xmm_result, xmm_key_first);               // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum <= XMM_REG_NUM_KEY_LAST - 1; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    aes_dec_key(xmm_result, xmm_temp, key, 0xb0);     // 192-bit key goes up to c0
+    aes_dec_key(xmm_result, xmm_temp, key, 0xc0);
+    __ aesdeclast(xmm_result, xmm_key_last);                    // xmm15 always came from key+0
+    __ pxor  (xmm_result, xmm_prev_block_cipher);               // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ movdqa(xmm_prev_block_cipher, xmm_prev_block_cipher_save);              // set up next r vector with cipher input from this block
+
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual,L_singleBlock_loopTop_192);
+    __ jmp(L_exit);
+
+    __ BIND(L_key_256);
+    // 256-bit code follows here (could be optimized to use parallelism)
+    __ movptr(pos, 0);
+    __ align(OptoLoopAlignment);
+    __ BIND(L_singleBlock_loopTop_256);
+    __ movdqu(xmm_result, Address(from, pos, Address::times_1, 0));   // get next 16 bytes of cipher input
+    __ movdqa(xmm_prev_block_cipher_save, xmm_result);              // save for next r vector
+    __ pxor  (xmm_result, xmm_key_first);               // do the aes dec rounds
+    for (int rnum = XMM_REG_NUM_KEY_FIRST + 1; rnum <= XMM_REG_NUM_KEY_LAST - 1; rnum++) {
+      __ aesdec(xmm_result, as_XMMRegister(rnum));
+    }
+    aes_dec_key(xmm_result, xmm_temp, key, 0xb0);     // 256-bit key goes up to e0
+    aes_dec_key(xmm_result, xmm_temp, key, 0xc0);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xd0);
+    aes_dec_key(xmm_result, xmm_temp, key, 0xe0);
+    __ aesdeclast(xmm_result, xmm_key_last);             // xmm15 came from key+0
+    __ pxor  (xmm_result, xmm_prev_block_cipher);               // xor with the current r vector
+    __ movdqu(Address(to, pos, Address::times_1, 0), xmm_result);     // store into the next 16 bytes of output
+    // no need to store r to memory until we exit
+    __ movdqa(xmm_prev_block_cipher, xmm_prev_block_cipher_save);              // set up next r vector with cipher input from this block
+
+    __ addptr(pos, AESBlockSize);
+    __ subptr(len_reg, AESBlockSize);
+    __ jcc(Assembler::notEqual,L_singleBlock_loopTop_256);
+    __ jmp(L_exit);
+
+    return start;
+  }
+
+
+
 #undef __
 #define __ masm->
 
@@ -3135,6 +3677,16 @@
     generate_arraycopy_stubs();
 
     generate_math_stubs();
+
+    // don't bother generating these AES intrinsic stubs unless global flag is set
+    if (UseAESIntrinsics) {
+      StubRoutines::x86::_key_shuffle_mask_addr = generate_key_shuffle_mask();  // needed by the others
+
+      StubRoutines::_aescrypt_encryptBlock = generate_aescrypt_encryptBlock();
+      StubRoutines::_aescrypt_decryptBlock = generate_aescrypt_decryptBlock();
+      StubRoutines::_cipherBlockChaining_encryptAESCrypt = generate_cipherBlockChaining_encryptAESCrypt();
+      StubRoutines::_cipherBlockChaining_decryptAESCrypt = generate_cipherBlockChaining_decryptAESCrypt_Parallel();
+    }
   }
 
  public:

--- a/src/cpu/x86/vm/stubRoutines_x86_32.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubRoutines_x86_32.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -44,3 +44,4 @@
 
 address StubRoutines::x86::_verify_mxcsr_entry         = NULL;
 address StubRoutines::x86::_verify_fpu_cntrl_wrd_entry = NULL;
+address StubRoutines::x86::_key_shuffle_mask_addr = NULL;

--- a/src/cpu/x86/vm/stubRoutines_x86_32.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubRoutines_x86_32.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -41,10 +41,14 @@
  private:
   static address _verify_mxcsr_entry;
   static address _verify_fpu_cntrl_wrd_entry;
+  // shuffle mask for fixing up 128-bit words consisting of big-endian 32-bit integers
+  static address _key_shuffle_mask_addr;
 
  public:
   static address verify_mxcsr_entry()                        { return _verify_mxcsr_entry; }
   static address verify_fpu_cntrl_wrd_entry()                { return _verify_fpu_cntrl_wrd_entry; }
+  static address key_shuffle_mask_addr()                     { return _key_shuffle_mask_addr; }
+
 };
 
   static bool    returns_to_call_stub(address return_pc)     { return return_pc == _call_stub_return_address; }

--- a/src/cpu/x86/vm/stubRoutines_x86_64.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubRoutines_x86_64.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -56,3 +56,4 @@
 address StubRoutines::x86::_double_sign_mask = NULL;
 address StubRoutines::x86::_double_sign_flip = NULL;
 address StubRoutines::x86::_mxcsr_std = NULL;
+address StubRoutines::x86::_key_shuffle_mask_addr = NULL;

--- a/src/cpu/x86/vm/stubRoutines_x86_64.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/stubRoutines_x86_64.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -54,6 +54,8 @@
   static address _double_sign_mask;
   static address _double_sign_flip;
   static address _mxcsr_std;
+  // shuffle mask for fixing up 128-bit words consisting of big-endian 32-bit integers
+  static address _key_shuffle_mask_addr;
 
  public:
 
@@ -116,6 +118,9 @@
   {
     return _mxcsr_std;
   }
+
+  static address key_shuffle_mask_addr()                     { return _key_shuffle_mask_addr; }
+
 };
 
 #endif // CPU_X86_VM_STUBROUTINES_X86_64_HPP

--- a/src/cpu/x86/vm/vm_version_x86.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/vm_version_x86.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -419,13 +419,16 @@
   if (UseAVX < 1)
     _cpuFeatures &= ~CPU_AVX;
 
+  if (!UseAES && !FLAG_IS_DEFAULT(UseAES))
+    _cpuFeatures &= ~CPU_AES;
+
   if (logical_processors_per_package() == 1) {
     // HT processor could be installed on a system which doesn't support HT.
     _cpuFeatures &= ~CPU_HT;
   }
 
   char buf[256];
-  jio_snprintf(buf, sizeof(buf), "(%u cores per cpu, %u threads per core) family %d model %d stepping %d%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s",
+  jio_snprintf(buf, sizeof(buf), "(%u cores per cpu, %u threads per core) family %d model %d stepping %d%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s",
                cores_per_cpu(), threads_per_core(),
                cpu_family(), _model, _stepping,
                (supports_cmov() ? ", cmov" : ""),
@@ -441,6 +444,7 @@
                (supports_popcnt() ? ", popcnt" : ""),
                (supports_avx()    ? ", avx" : ""),
                (supports_avx2()   ? ", avx2" : ""),
+               (supports_aes()    ? ", aes" : ""),
                (supports_mmx_ext() ? ", mmxext" : ""),
                (supports_3dnow_prefetch() ? ", 3dnowpref" : ""),
                (supports_lzcnt()   ? ", lzcnt": ""),
@@ -472,6 +476,29 @@
   if (!supports_avx ()) // Drop to 0 if no AVX  support
     UseAVX = 0;
 
+  // Use AES instructions if available.
+  if (supports_aes()) {
+    if (FLAG_IS_DEFAULT(UseAES)) {
+      UseAES = true;
+    }
+  } else if (UseAES) {
+    if (!FLAG_IS_DEFAULT(UseAES))
+      warning("AES instructions not available on this CPU");
+    FLAG_SET_DEFAULT(UseAES, false);
+  }
+
+  // The AES intrinsic stubs require AES instruction support (of course)
+  // but also require AVX mode for misaligned SSE access
+  if (UseAES && (UseAVX > 0)) {
+    if (FLAG_IS_DEFAULT(UseAESIntrinsics)) {
+      UseAESIntrinsics = true;
+    }
+  } else if (UseAESIntrinsics) {
+    if (!FLAG_IS_DEFAULT(UseAESIntrinsics))
+      warning("AES intrinsics not available on this CPU");
+    FLAG_SET_DEFAULT(UseAESIntrinsics, false);
+  }
+
 #ifdef COMPILER2
   if (UseFPUForSpilling) {
     if (UseSSE < 2) {
@@ -714,6 +741,9 @@
     if (UseAVX > 0) {
       tty->print("  UseAVX=%d",UseAVX);
     }
+    if (UseAES) {
+      tty->print("  UseAES=1");
+    }
     tty->cr();
     tty->print("Allocation");
     if (AllocatePrefetchStyle <= 0 || UseSSE == 0 && !supports_3dnow_prefetch()) {

--- a/src/cpu/x86/vm/vm_version_x86.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/vm_version_x86.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -78,7 +78,9 @@
                sse4_2   : 1,
                         : 2,
                popcnt   : 1,
-                        : 3,
+                        : 1,
+               aes      : 1,
+                        : 1,
                osxsave  : 1,
                avx      : 1,
                         : 3;
@@ -244,7 +246,8 @@
     CPU_TSC    = (1 << 15),
     CPU_TSCINV = (1 << 16),
     CPU_AVX    = (1 << 17),
-    CPU_AVX2   = (1 << 18)
+    CPU_AVX2   = (1 << 18),
+    CPU_AES    = (1 << 19)
   } cpuFeatureFlags;
 
   enum {
@@ -420,6 +423,8 @@
       result |= CPU_TSC;
     if (_cpuid_info.ext_cpuid7_edx.bits.tsc_invariance != 0)
       result |= CPU_TSCINV;
+    if (_cpuid_info.std_cpuid1_ecx.bits.aes != 0)
+      result |= CPU_AES;
 
     // AMD features.
     if (is_amd()) {
@@ -544,6 +549,7 @@
   static bool supports_avx()      { return (_cpuFeatures & CPU_AVX) != 0; }
   static bool supports_avx2()     { return (_cpuFeatures & CPU_AVX2) != 0; }
   static bool supports_tsc()      { return (_cpuFeatures & CPU_TSC)    != 0; }
+  static bool supports_aes()      { return (_cpuFeatures & CPU_AES) != 0; }
 
   // Intel features
   static bool is_intel_family_core() { return is_intel() &&

--- a/src/cpu/x86/vm/x86.ad	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/x86/vm/x86.ad	Mon Nov 05 13:55:31 2012 -0800
@@ -4102,9 +4102,158 @@
 
 // ----------------------- LogicalRightShift -----------------------------------
 
-// Shorts/Chars vector logical right shift produces incorrect Java result
+// Shorts vector logical right shift produces incorrect Java result
 // for negative data because java code convert short value into int with
-// sign extension before a shift.
+// sign extension before a shift. But char vectors are fine since chars are
+// unsigned values.
+
+instruct vsrl2S(vecS dst, vecS shift) %{
+  predicate(n->as_Vector()->length() == 2);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed2S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, $shift$$XMMRegister);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl2S_imm(vecS dst, immI8 shift) %{
+  predicate(n->as_Vector()->length() == 2);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed2S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, (int)$shift$$constant);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl2S_reg(vecS dst, vecS src, vecS shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 2);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed2S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, $shift$$XMMRegister, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl2S_reg_imm(vecS dst, vecS src, immI8 shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 2);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed2S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, (int)$shift$$constant, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl4S(vecD dst, vecS shift) %{
+  predicate(n->as_Vector()->length() == 4);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed4S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, $shift$$XMMRegister);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl4S_imm(vecD dst, immI8 shift) %{
+  predicate(n->as_Vector()->length() == 4);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed4S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, (int)$shift$$constant);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl4S_reg(vecD dst, vecD src, vecS shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 4);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed4S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, $shift$$XMMRegister, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl4S_reg_imm(vecD dst, vecD src, immI8 shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 4);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed4S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, (int)$shift$$constant, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl8S(vecX dst, vecS shift) %{
+  predicate(n->as_Vector()->length() == 8);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed8S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, $shift$$XMMRegister);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl8S_imm(vecX dst, immI8 shift) %{
+  predicate(n->as_Vector()->length() == 8);
+  match(Set dst (URShiftVS dst shift));
+  format %{ "psrlw   $dst,$shift\t! logical right shift packed8S" %}
+  ins_encode %{
+    __ psrlw($dst$$XMMRegister, (int)$shift$$constant);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl8S_reg(vecX dst, vecX src, vecS shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 8);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed8S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, $shift$$XMMRegister, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl8S_reg_imm(vecX dst, vecX src, immI8 shift) %{
+  predicate(UseAVX > 0 && n->as_Vector()->length() == 8);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed8S" %}
+  ins_encode %{
+    bool vector256 = false;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, (int)$shift$$constant, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl16S_reg(vecY dst, vecY src, vecS shift) %{
+  predicate(UseAVX > 1 && n->as_Vector()->length() == 16);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed16S" %}
+  ins_encode %{
+    bool vector256 = true;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, $shift$$XMMRegister, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
+
+instruct vsrl16S_reg_imm(vecY dst, vecY src, immI8 shift) %{
+  predicate(UseAVX > 1 && n->as_Vector()->length() == 16);
+  match(Set dst (URShiftVS src shift));
+  format %{ "vpsrlw  $dst,$src,$shift\t! logical right shift packed16S" %}
+  ins_encode %{
+    bool vector256 = true;
+    __ vpsrlw($dst$$XMMRegister, $src$$XMMRegister, (int)$shift$$constant, vector256);
+  %}
+  ins_pipe( pipe_slow );
+%}
 
 // Integers vector logical right shift
 instruct vsrl2I(vecD dst, vecS shift) %{

--- a/src/cpu/zero/vm/cppInterpreterGenerator_zero.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/cppInterpreterGenerator_zero.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -31,12 +31,17 @@
     return _masm;
   }
 
- protected:
-  address generate_entry(address entry_point) {
-    ZeroEntry *entry = (ZeroEntry *) assembler()->pc();
-    assembler()->advance(sizeof(ZeroEntry));
+ public:
+  static address generate_entry_impl(MacroAssembler* masm, address entry_point) {
+    ZeroEntry *entry = (ZeroEntry *) masm->pc();
+    masm->advance(sizeof(ZeroEntry));
     entry->set_entry_point(entry_point);
     return (address) entry;
   }
 
+ protected:
+  address generate_entry(address entry_point) {
+        return generate_entry_impl(assembler(), entry_point);
+  }
+
 #endif // CPU_ZERO_VM_CPPINTERPRETERGENERATOR_ZERO_HPP

--- a/src/cpu/zero/vm/cppInterpreter_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/cppInterpreter_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -180,25 +180,6 @@
         method, istate->osr_entry(), istate->osr_buf(), THREAD);
       return;
     }
-    else if (istate->msg() == BytecodeInterpreter::call_method_handle) {
-      oop method_handle = istate->callee();
-
-      // Trim back the stack to put the parameters at the top
-      stack->set_sp(istate->stack() + 1);
-
-      // Make the call
-      process_method_handle(method_handle, THREAD);
-      fixup_after_potential_safepoint();
-
-      // Convert the result
-      istate->set_stack(stack->sp() - 1);
-
-      // Restore the stack
-      stack->set_sp(istate->stack_limit() + 1);
-
-      // Resume the interpreter
-      istate->set_msg(BytecodeInterpreter::method_resume);
-    }
     else {
       ShouldNotReachHere();
     }
@@ -535,35 +516,35 @@
   if (entry->is_volatile()) {
     switch (entry->flag_state()) {
     case ctos:
-      SET_LOCALS_INT(object->char_field_acquire(entry->f2()), 0);
+      SET_LOCALS_INT(object->char_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case btos:
-      SET_LOCALS_INT(object->byte_field_acquire(entry->f2()), 0);
+      SET_LOCALS_INT(object->byte_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case stos:
-      SET_LOCALS_INT(object->short_field_acquire(entry->f2()), 0);
+      SET_LOCALS_INT(object->short_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case itos:
-      SET_LOCALS_INT(object->int_field_acquire(entry->f2()), 0);
+      SET_LOCALS_INT(object->int_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case ltos:
-      SET_LOCALS_LONG(object->long_field_acquire(entry->f2()), 0);
+      SET_LOCALS_LONG(object->long_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case ftos:
-      SET_LOCALS_FLOAT(object->float_field_acquire(entry->f2()), 0);
+      SET_LOCALS_FLOAT(object->float_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case dtos:
-      SET_LOCALS_DOUBLE(object->double_field_acquire(entry->f2()), 0);
+      SET_LOCALS_DOUBLE(object->double_field_acquire(entry->f2_as_index()), 0);
       break;
 
     case atos:
-      SET_LOCALS_OBJECT(object->obj_field_acquire(entry->f2()), 0);
+      SET_LOCALS_OBJECT(object->obj_field_acquire(entry->f2_as_index()), 0);
       break;
 
     default:
@@ -573,35 +554,35 @@
   else {
     switch (entry->flag_state()) {
     case ctos:
-      SET_LOCALS_INT(object->char_field(entry->f2()), 0);
+      SET_LOCALS_INT(object->char_field(entry->f2_as_index()), 0);
       break;
 
     case btos:
-      SET_LOCALS_INT(object->byte_field(entry->f2()), 0);
+      SET_LOCALS_INT(object->byte_field(entry->f2_as_index()), 0);
       break;
 
     case stos:
-      SET_LOCALS_INT(object->short_field(entry->f2()), 0);
+      SET_LOCALS_INT(object->short_field(entry->f2_as_index()), 0);
       break;
 
     case itos:
-      SET_LOCALS_INT(object->int_field(entry->f2()), 0);
+      SET_LOCALS_INT(object->int_field(entry->f2_as_index()), 0);
       break;
 
     case ltos:
-      SET_LOCALS_LONG(object->long_field(entry->f2()), 0);
+      SET_LOCALS_LONG(object->long_field(entry->f2_as_index()), 0);
       break;
 
     case ftos:
-      SET_LOCALS_FLOAT(object->float_field(entry->f2()), 0);
+      SET_LOCALS_FLOAT(object->float_field(entry->f2_as_index()), 0);
       break;
 
     case dtos:
-      SET_LOCALS_DOUBLE(object->double_field(entry->f2()), 0);
+      SET_LOCALS_DOUBLE(object->double_field(entry->f2_as_index()), 0);
       break;
 
     case atos:
-      SET_LOCALS_OBJECT(object->obj_field(entry->f2()), 0);
+      SET_LOCALS_OBJECT(object->obj_field(entry->f2_as_index()), 0);
       break;
 
     default:
@@ -629,516 +610,6 @@
   return 0;
 }
 
-int CppInterpreter::method_handle_entry(Method* method,
-                                        intptr_t UNUSED, TRAPS) {
-  JavaThread *thread = (JavaThread *) THREAD;
-  ZeroStack *stack = thread->zero_stack();
-  int argument_slots = method->size_of_parameters();
-  int result_slots = type2size[result_type_of(method)];
-  intptr_t *vmslots = stack->sp();
-  intptr_t *unwind_sp = vmslots + argument_slots;
-
-  // Find the MethodType
-  address p = (address) method;
-  for (jint* pc = method->method_type_offsets_chain(); (*pc) != -1; pc++) {
-    p = *(address*)(p + (*pc));
-  }
-  oop method_type = (oop) p;
-
-  // The MethodHandle is in the slot after the arguments
-  int num_vmslots = argument_slots - 1;
-  oop method_handle = VMSLOTS_OBJECT(num_vmslots);
-
-  // InvokeGeneric requires some extra shuffling
-  oop mhtype = java_lang_invoke_MethodHandle::type(method_handle);
-  bool is_exact = mhtype == method_type;
-  if (!is_exact) {
-    if (true || // FIXME
-        method->intrinsic_id() == vmIntrinsics::_invokeExact) {
-      CALL_VM_NOCHECK_NOFIX(
-        SharedRuntime::throw_WrongMethodTypeException(
-          thread, method_type, mhtype));
-      // NB all oops trashed!
-      assert(HAS_PENDING_EXCEPTION, "should do");
-      stack->set_sp(unwind_sp);
-      return 0;
-    }
-    assert(method->intrinsic_id() == vmIntrinsics::_invokeGeneric, "should be");
-
-    // Load up an adapter from the calling type
-    // NB the x86 code for this (in methodHandles_x86.cpp, search for
-    // "genericInvoker") is really really odd.  I'm hoping it's trying
-    // to accomodate odd VM/class library combinations I can ignore.
-    oop adapter = NULL; //FIXME: load the adapter from the CP cache
-    IF (adapter == NULL) {
-      CALL_VM_NOCHECK_NOFIX(
-        SharedRuntime::throw_WrongMethodTypeException(
-          thread, method_type, mhtype));
-      // NB all oops trashed!
-      assert(HAS_PENDING_EXCEPTION, "should do");
-      stack->set_sp(unwind_sp);
-      return 0;
-    }
-
-    // Adapters are shared among form-families of method-type.  The
-    // type being called is passed as a trusted first argument so that
-    // the adapter knows the actual types of its arguments and return
-    // values.
-    insert_vmslots(num_vmslots + 1, 1, THREAD);
-    if (HAS_PENDING_EXCEPTION) {
-      // NB all oops trashed!
-      stack->set_sp(unwind_sp);
-      return 0;
-    }
-
-    vmslots = stack->sp();
-    num_vmslots++;
-    SET_VMSLOTS_OBJECT(method_type, num_vmslots);
-
-    method_handle = adapter;
-  }
-
-  // Start processing
-  process_method_handle(method_handle, THREAD);
-  if (HAS_PENDING_EXCEPTION)
-    result_slots = 0;
-
-  // If this is an invokeExact then the eventual callee will not
-  // have unwound the method handle argument so we have to do it.
-  // If a result is being returned the it will be above the method
-  // handle argument we're unwinding.
-  if (is_exact) {
-    intptr_t result[2];
-    for (int i = 0; i < result_slots; i++)
-      result[i] = stack->pop();
-    stack->pop();
-    for (int i = result_slots - 1; i >= 0; i--)
-      stack->push(result[i]);
-  }
-
-  // Check
-  assert(stack->sp() == unwind_sp - result_slots, "should be");
-
-  // No deoptimized frames on the stack
-  return 0;
-}
-
-void CppInterpreter::process_method_handle(oop method_handle, TRAPS) {
-  JavaThread *thread = (JavaThread *) THREAD;
-  ZeroStack *stack = thread->zero_stack();
-  intptr_t *vmslots = stack->sp();
-
-  bool direct_to_method = false;
-  BasicType src_rtype = T_ILLEGAL;
-  BasicType dst_rtype = T_ILLEGAL;
-
-  MethodHandleEntry *entry =
-    java_lang_invoke_MethodHandle::vmentry(method_handle);
-  MethodHandles::EntryKind entry_kind =
-    (MethodHandles::EntryKind) (((intptr_t) entry) & 0xffffffff);
-
-  Method* method = NULL;
-  switch (entry_kind) {
-  case MethodHandles::_invokestatic_mh:
-    direct_to_method = true;
-    break;
-
-  case MethodHandles::_invokespecial_mh:
-  case MethodHandles::_invokevirtual_mh:
-  case MethodHandles::_invokeinterface_mh:
-    {
-      oop receiver =
-        VMSLOTS_OBJECT(
-          java_lang_invoke_MethodHandle::vmslots(method_handle) - 1);
-      if (receiver == NULL) {
-          stack->set_sp(calculate_unwind_sp(stack, method_handle));
-          CALL_VM_NOCHECK_NOFIX(
-            throw_exception(
-              thread, vmSymbols::java_lang_NullPointerException()));
-          // NB all oops trashed!
-          assert(HAS_PENDING_EXCEPTION, "should do");
-          return;
-      }
-      if (entry_kind != MethodHandles::_invokespecial_mh) {
-        intptr_t index = java_lang_invoke_DirectMethodHandle::vmindex(method_handle);
-        InstanceKlass* rcvrKlass =
-          (InstanceKlass *) receiver->klass();
-        if (entry_kind == MethodHandles::_invokevirtual_mh) {
-          method = (Method*) rcvrKlass->start_of_vtable()[index];
-        }
-        else {
-          oop iclass = java_lang_invoke_MethodHandle::next_target(method_handle);
-          itableOffsetEntry* ki =
-            (itableOffsetEntry *) rcvrKlass->start_of_itable();
-          int i, length = rcvrKlass->itable_length();
-          for (i = 0; i < length; i++, ki++ ) {
-            if (ki->interface_klass() == iclass)
-              break;
-          }
-          if (i == length) {
-            stack->set_sp(calculate_unwind_sp(stack, method_handle));
-            CALL_VM_NOCHECK_NOFIX(
-              throw_exception(
-                thread, vmSymbols::java_lang_IncompatibleClassChangeError()));
-            // NB all oops trashed!
-            assert(HAS_PENDING_EXCEPTION, "should do");
-            return;
-          }
-          itableMethodEntry* im = ki->first_method_entry(receiver->klass());
-          method = im[index].method();
-          if (method == NULL) {
-            stack->set_sp(calculate_unwind_sp(stack, method_handle));
-            CALL_VM_NOCHECK_NOFIX(
-              throw_exception(
-                thread, vmSymbols::java_lang_AbstractMethodError()));
-            // NB all oops trashed!
-            assert(HAS_PENDING_EXCEPTION, "should do");
-            return;
-          }
-        }
-      }
-    }
-    direct_to_method = true;
-    break;
-
-  case MethodHandles::_bound_ref_direct_mh:
-  case MethodHandles::_bound_int_direct_mh:
-  case MethodHandles::_bound_long_direct_mh:
-    direct_to_method = true;
-    // fall through
-  case MethodHandles::_bound_ref_mh:
-  case MethodHandles::_bound_int_mh:
-  case MethodHandles::_bound_long_mh:
-    {
-      BasicType arg_type  = T_ILLEGAL;
-      int       arg_mask  = -1;
-      int       arg_slots = -1;
-      MethodHandles::get_ek_bound_mh_info(
-        entry_kind, arg_type, arg_mask, arg_slots);
-      int arg_slot =
-        java_lang_invoke_BoundMethodHandle::vmargslot(method_handle);
-
-      // Create the new slot(s)
-      intptr_t *unwind_sp = calculate_unwind_sp(stack, method_handle);
-      insert_vmslots(arg_slot, arg_slots, THREAD);
-      if (HAS_PENDING_EXCEPTION) {
-        // all oops trashed
-        stack->set_sp(unwind_sp);
-        return;
-      }
-      vmslots = stack->sp();
-
-      // Store bound argument into new stack slot
-      oop arg = java_lang_invoke_BoundMethodHandle::argument(method_handle);
-      if (arg_type == T_OBJECT) {
-        assert(arg_slots == 1, "should be");
-        SET_VMSLOTS_OBJECT(arg, arg_slot);
-      }
-      else {
-        jvalue arg_value;
-        arg_type = java_lang_boxing_object::get_value(arg, &arg_value);
-        switch (arg_type) {
-        case T_BOOLEAN:
-          SET_VMSLOTS_INT(arg_value.z, arg_slot);
-          break;
-        case T_CHAR:
-          SET_VMSLOTS_INT(arg_value.c, arg_slot);
-          break;
-        case T_BYTE:
-          SET_VMSLOTS_INT(arg_value.b, arg_slot);
-          break;
-        case T_SHORT:
-          SET_VMSLOTS_INT(arg_value.s, arg_slot);
-          break;
-        case T_INT:
-          SET_VMSLOTS_INT(arg_value.i, arg_slot);
-          break;
-        case T_FLOAT:
-          SET_VMSLOTS_FLOAT(arg_value.f, arg_slot);
-          break;
-        case T_LONG:
-          SET_VMSLOTS_LONG(arg_value.j, arg_slot + 1);
-          break;
-        case T_DOUBLE:
-          SET_VMSLOTS_DOUBLE(arg_value.d, arg_slot + 1);
-          break;
-        default:
-          tty->print_cr("unhandled type %s", type2name(arg_type));
-          ShouldNotReachHere();
-        }
-      }
-    }
-    break;
-
-  case MethodHandles::_adapter_retype_only:
-  case MethodHandles::_adapter_retype_raw:
-    src_rtype = result_type_of_handle(
-      java_lang_invoke_MethodHandle::next_target(method_handle));
-    dst_rtype = result_type_of_handle(method_handle);
-    break;
-
-  case MethodHandles::_adapter_check_cast:
-    {
-      int arg_slot =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      oop arg = VMSLOTS_OBJECT(arg_slot);
-      if (arg != NULL) {
-        Klass* objKlassOop = arg->klass();
-        Klass* klassOf = java_lang_Class::as_Klass(
-          java_lang_invoke_AdapterMethodHandle::argument(method_handle));
-
-        if (objKlassOop != klassOf &&
-            !objKlassOop->is_subtype_of(klassOf)) {
-          ResourceMark rm(THREAD);
-          const char* objName = Klass::cast(objKlassOop)->external_name();
-          const char* klassName = Klass::cast(klassOf)->external_name();
-          char* message = SharedRuntime::generate_class_cast_message(
-            objName, klassName);
-
-          stack->set_sp(calculate_unwind_sp(stack, method_handle));
-          CALL_VM_NOCHECK_NOFIX(
-            throw_exception(
-              thread, vmSymbols::java_lang_ClassCastException(), message));
-          // NB all oops trashed!
-          assert(HAS_PENDING_EXCEPTION, "should do");
-          return;
-        }
-      }
-    }
-    break;
-
-  case MethodHandles::_adapter_dup_args:
-    {
-      int arg_slot =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      int conv =
-        java_lang_invoke_AdapterMethodHandle::conversion(method_handle);
-      int num_slots = -MethodHandles::adapter_conversion_stack_move(conv);
-      assert(num_slots > 0, "should be");
-
-      // Create the new slot(s)
-      intptr_t *unwind_sp = calculate_unwind_sp(stack, method_handle);
-      stack->overflow_check(num_slots, THREAD);
-      if (HAS_PENDING_EXCEPTION) {
-        // all oops trashed
-        stack->set_sp(unwind_sp);
-        return;
-      }
-
-      // Duplicate the arguments
-      for (int i = num_slots - 1; i >= 0; i--)
-        stack->push(*VMSLOTS_SLOT(arg_slot + i));
-
-      vmslots = stack->sp(); // unused, but let the compiler figure that out
-    }
-    break;
-
-  case MethodHandles::_adapter_drop_args:
-    {
-      int arg_slot =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      int conv =
-        java_lang_invoke_AdapterMethodHandle::conversion(method_handle);
-      int num_slots = MethodHandles::adapter_conversion_stack_move(conv);
-      assert(num_slots > 0, "should be");
-
-      remove_vmslots(arg_slot, num_slots, THREAD); // doesn't trap
-      vmslots = stack->sp(); // unused, but let the compiler figure that out
-    }
-    break;
-
-  case MethodHandles::_adapter_opt_swap_1:
-  case MethodHandles::_adapter_opt_swap_2:
-  case MethodHandles::_adapter_opt_rot_1_up:
-  case MethodHandles::_adapter_opt_rot_1_down:
-  case MethodHandles::_adapter_opt_rot_2_up:
-  case MethodHandles::_adapter_opt_rot_2_down:
-    {
-      int arg1 =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      int conv =
-        java_lang_invoke_AdapterMethodHandle::conversion(method_handle);
-      int arg2 = MethodHandles::adapter_conversion_vminfo(conv);
-
-      int swap_bytes = 0, rotate = 0;
-      MethodHandles::get_ek_adapter_opt_swap_rot_info(
-        entry_kind, swap_bytes, rotate);
-      int swap_slots = swap_bytes >> LogBytesPerWord;
-
-      intptr_t tmp;
-      switch (rotate) {
-      case 0: // swap
-        for (int i = 0; i < swap_slots; i++) {
-          tmp = *VMSLOTS_SLOT(arg1 + i);
-          SET_VMSLOTS_SLOT(VMSLOTS_SLOT(arg2 + i), arg1 + i);
-          SET_VMSLOTS_SLOT(&tmp, arg2 + i);
-        }
-        break;
-
-      case 1: // up
-        assert(arg1 - swap_slots > arg2, "should be");
-
-        tmp = *VMSLOTS_SLOT(arg1);
-        for (int i = arg1 - swap_slots; i >= arg2; i--)
-          SET_VMSLOTS_SLOT(VMSLOTS_SLOT(i), i + swap_slots);
-        SET_VMSLOTS_SLOT(&tmp, arg2);
-
-        break;
-
-      case -1: // down
-        assert(arg2 - swap_slots > arg1, "should be");
-
-        tmp = *VMSLOTS_SLOT(arg1);
-        for (int i = arg1 + swap_slots; i <= arg2; i++)
-          SET_VMSLOTS_SLOT(VMSLOTS_SLOT(i), i - swap_slots);
-        SET_VMSLOTS_SLOT(&tmp, arg2);
-        break;
-
-      default:
-        ShouldNotReachHere();
-      }
-    }
-    break;
-
-  case MethodHandles::_adapter_opt_i2l:
-    {
-      int arg_slot =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      int arg = VMSLOTS_INT(arg_slot);
-      intptr_t *unwind_sp = calculate_unwind_sp(stack, method_handle);
-      insert_vmslots(arg_slot, 1, THREAD);
-      if (HAS_PENDING_EXCEPTION) {
-        // all oops trashed
-        stack->set_sp(unwind_sp);
-        return;
-      }
-      vmslots = stack->sp();
-      arg_slot++;
-      SET_VMSLOTS_LONG(arg, arg_slot);
-    }
-    break;
-
-  case MethodHandles::_adapter_opt_unboxi:
-  case MethodHandles::_adapter_opt_unboxl:
-    {
-      int arg_slot =
-        java_lang_invoke_AdapterMethodHandle::vmargslot(method_handle);
-      oop arg = VMSLOTS_OBJECT(arg_slot);
-      jvalue arg_value;
-      if (arg == NULL) {
-        // queue a nullpointer exception for the caller
-        stack->set_sp(calculate_unwind_sp(stack, method_handle));
-        CALL_VM_NOCHECK_NOFIX(
-          throw_exception(
-            thread, vmSymbols::java_lang_NullPointerException()));
-        // NB all oops trashed!
-        assert(HAS_PENDING_EXCEPTION, "should do");
-        return;
-      }
-      BasicType arg_type = java_lang_boxing_object::get_value(arg, &arg_value);
-      if (arg_type == T_LONG || arg_type == T_DOUBLE) {
-        intptr_t *unwind_sp = calculate_unwind_sp(stack, method_handle);
-        insert_vmslots(arg_slot, 1, THREAD);
-        if (HAS_PENDING_EXCEPTION) {
-          // all oops trashed
-          stack->set_sp(unwind_sp);
-          return;
-        }
-        vmslots = stack->sp();
-        arg_slot++;
-      }
-      switch (arg_type) {
-      case T_BOOLEAN:
-        SET_VMSLOTS_INT(arg_value.z, arg_slot);
-        break;
-      case T_CHAR:
-        SET_VMSLOTS_INT(arg_value.c, arg_slot);
-        break;
-      case T_BYTE:
-        SET_VMSLOTS_INT(arg_value.b, arg_slot);
-        break;
-      case T_SHORT:
-        SET_VMSLOTS_INT(arg_value.s, arg_slot);
-        break;
-      case T_INT:
-        SET_VMSLOTS_INT(arg_value.i, arg_slot);
-        break;
-      case T_FLOAT:
-        SET_VMSLOTS_FLOAT(arg_value.f, arg_slot);
-        break;
-      case T_LONG:
-        SET_VMSLOTS_LONG(arg_value.j, arg_slot);
-        break;
-      case T_DOUBLE:
-        SET_VMSLOTS_DOUBLE(arg_value.d, arg_slot);
-        break;
-      default:
-        tty->print_cr("unhandled type %s", type2name(arg_type));
-        ShouldNotReachHere();
-      }
-    }
-    break;
-
-  default:
-    tty->print_cr("unhandled entry_kind %s",
-                  MethodHandles::entry_name(entry_kind));
-    ShouldNotReachHere();
-  }
-
-  // Continue along the chain
-  if (direct_to_method) {
-    if (method == NULL) {
-      method =
-        (Method*) java_lang_invoke_MethodHandle::vmtarget(method_handle);
-    }
-    address entry_point = method->from_interpreted_entry();
-    Interpreter::invoke_method(method, entry_point, THREAD);
-  }
-  else {
-    process_method_handle(
-      java_lang_invoke_MethodHandle::next_target(method_handle), THREAD);
-  }
-  // NB all oops now trashed
-
-  // Adapt the result type, if necessary
-  if (src_rtype != dst_rtype && !HAS_PENDING_EXCEPTION) {
-    switch (dst_rtype) {
-    case T_VOID:
-      for (int i = 0; i < type2size[src_rtype]; i++)
-        stack->pop();
-      return;
-
-    case T_INT:
-      switch (src_rtype) {
-      case T_VOID:
-        stack->overflow_check(1, CHECK);
-        stack->push(0);
-        return;
-
-      case T_BOOLEAN:
-      case T_CHAR:
-      case T_BYTE:
-      case T_SHORT:
-        return;
-      }
-      // INT results sometimes need narrowing
-    case T_BOOLEAN:
-    case T_CHAR:
-    case T_BYTE:
-    case T_SHORT:
-      switch (src_rtype) {
-      case T_INT:
-        return;
-      }
-    }
-
-    tty->print_cr("unhandled conversion:");
-    tty->print_cr("src_rtype = %s", type2name(src_rtype));
-    tty->print_cr("dst_rtype = %s", type2name(dst_rtype));
-    ShouldNotReachHere();
-  }
-}
-
 // The new slots will be inserted before slot insert_before.
 // Slots < insert_before will have the same slot number after the insert.
 // Slots >= insert_before will become old_slot + num_slots.
@@ -1380,10 +851,6 @@
     entry_point = ((InterpreterGenerator*) this)->generate_abstract_entry();
     break;
 
-  case Interpreter::method_handle:
-    entry_point = ((InterpreterGenerator*) this)->generate_method_handle_entry();
-    break;
-
   case Interpreter::java_lang_math_sin:
   case Interpreter::java_lang_math_cos:
   case Interpreter::java_lang_math_tan:
@@ -1391,6 +858,8 @@
   case Interpreter::java_lang_math_log:
   case Interpreter::java_lang_math_log10:
   case Interpreter::java_lang_math_sqrt:
+  case Interpreter::java_lang_math_pow:
+  case Interpreter::java_lang_math_exp:
     entry_point = ((InterpreterGenerator*) this)->generate_math_entry(kind);
     break;
 

--- a/src/cpu/zero/vm/cppInterpreter_zero.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/cppInterpreter_zero.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -36,7 +36,6 @@
   static int native_entry(Method* method, intptr_t UNUSED, TRAPS);
   static int accessor_entry(Method* method, intptr_t UNUSED, TRAPS);
   static int empty_entry(Method* method, intptr_t UNUSED, TRAPS);
-  static int method_handle_entry(Method* method, intptr_t UNUSED, TRAPS);
 
  public:
   // Main loop of normal_entry
@@ -44,7 +43,6 @@
 
  private:
   // Helpers for method_handle_entry
-  static void process_method_handle(oop method_handle, TRAPS);
   static void insert_vmslots(int insert_before, int num_slots, TRAPS);
   static void remove_vmslots(int first_slot, int num_slots, TRAPS);
   static BasicType result_type_of_handle(oop method_handle);

--- a/src/cpu/zero/vm/frame_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/frame_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -351,7 +351,7 @@
   switch (offset) {
   case pc_off:
     strncpy(fieldbuf, "pc", buflen);
-    if (method()->is_oop()) {
+    if (method()->is_method()) {
       nmethod *code = method()->code();
       if (code && code->pc_desc_at(pc())) {
         SimpleScopeDesc ssd(code, pc());
@@ -367,7 +367,7 @@
 
   case method_off:
     strncpy(fieldbuf, "method", buflen);
-    if (method()->is_oop()) {
+    if (method()->is_method()) {
       method()->name_and_sig_as_C_string(valuebuf, buflen);
     }
     return;
@@ -378,7 +378,7 @@
   }
 
   // Variable part
-  if (method()->is_oop()) {
+  if (method()->is_method()) {
     identify_vp_word(frame_index, addr_of_word(offset),
                      addr_of_word(header_words + 1),
                      unextended_sp() + method()->max_stack(),
@@ -430,4 +430,3 @@
   // unused... but returns fp() to minimize changes introduced by 7087445
   return fp();
 }
-

--- a/src/cpu/zero/vm/frame_zero.inline.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/frame_zero.inline.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -36,6 +36,8 @@
   _deopt_state = unknown;
 }
 
+inline address  frame::sender_pc()           const { ShouldNotCallThis();  }
+
 inline frame::frame(ZeroFrame* zf, intptr_t* sp) {
   _zeroframe = zf;
   _sp = sp;

--- a/src/cpu/zero/vm/icBuffer_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/icBuffer_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -40,7 +40,7 @@
 }
 
 void InlineCacheBuffer::assemble_ic_buffer_code(address code_begin,
-                                                Metadata* cached_oop,
+                                                void* cached_oop,
                                                 address entry_point) {
   // NB ic_stub_code_size() must return the size of the code we generate
   ShouldNotCallThis();
@@ -51,7 +51,6 @@
   ShouldNotCallThis();
 }
 
-Metadata* InlineCacheBuffer::ic_buffer_cached_oop(address code_begin) {
-  // NB ic_stub_code_size() must return the size of the code we generate
+void* InlineCacheBuffer::ic_buffer_cached_value(address code_begin) {
   ShouldNotCallThis();
 }

--- a/src/cpu/zero/vm/methodHandles_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/methodHandles_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -24,26 +24,159 @@
  */
 
 #include "precompiled.hpp"
+#include "interpreter/interpreterGenerator.hpp"
 #include "interpreter/interpreter.hpp"
 #include "memory/allocation.inline.hpp"
 #include "prims/methodHandles.hpp"
 
-int MethodHandles::adapter_conversion_ops_supported_mask() {
-  return ((1<<java_lang_invoke_AdapterMethodHandle::OP_RETYPE_ONLY)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_RETYPE_RAW)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_CHECK_CAST)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_PRIM_TO_PRIM)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_REF_TO_PRIM)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_SWAP_ARGS)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_ROT_ARGS)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_DUP_ARGS)
-         |(1<<java_lang_invoke_AdapterMethodHandle::OP_DROP_ARGS)
-         //|(1<<java_lang_invoke_AdapterMethodHandle::OP_SPREAD_ARGS) //BUG!
-         );
-  // FIXME: MethodHandlesTest gets a crash if we enable OP_SPREAD_ARGS.
+void MethodHandles::invoke_target(Method* method, TRAPS) {
+
+  JavaThread *thread = (JavaThread *) THREAD;
+  ZeroStack *stack = thread->zero_stack();
+  InterpreterFrame *frame = thread->top_zero_frame()->as_interpreter_frame();
+  interpreterState istate = frame->interpreter_state();
+
+  // Trim back the stack to put the parameters at the top
+  stack->set_sp(istate->stack() + 1);
+
+  Interpreter::invoke_method(method, method->from_interpreted_entry(), THREAD);
+
+  // Convert the result
+  istate->set_stack(stack->sp() - 1);
+
+}
+
+oop MethodHandles::popFromStack(TRAPS) {
+
+  JavaThread *thread = (JavaThread *) THREAD;
+  InterpreterFrame *frame = thread->top_zero_frame()->as_interpreter_frame();
+  interpreterState istate = frame->interpreter_state();
+  intptr_t* topOfStack = istate->stack();
+
+  oop top = STACK_OBJECT(-1);
+  MORE_STACK(-1);
+  istate->set_stack(topOfStack);
+
+  return top;
+
+}
+
+int MethodHandles::method_handle_entry_invokeBasic(Method* method, intptr_t UNUSED, TRAPS) {
+
+  JavaThread *thread = (JavaThread *) THREAD;
+  InterpreterFrame *frame = thread->top_zero_frame()->as_interpreter_frame();
+  interpreterState istate = frame->interpreter_state();
+  intptr_t* topOfStack = istate->stack();
+
+  // 'this' is a MethodHandle. We resolve the target method by accessing this.form.vmentry.vmtarget.
+  int numArgs = method->size_of_parameters();
+  oop lform1 = java_lang_invoke_MethodHandle::form(STACK_OBJECT(-numArgs)); // this.form
+  oop vmEntry1 = java_lang_invoke_LambdaForm::vmentry(lform1);
+  Method* vmtarget = (Method*) java_lang_invoke_MemberName::vmtarget(vmEntry1);
+
+  invoke_target(vmtarget, THREAD);
+
+  // No deoptimized frames on the stack
+  return 0;
+}
+
+int MethodHandles::method_handle_entry_linkToStaticOrSpecial(Method* method, intptr_t UNUSED, TRAPS) {
+
+  // Pop appendix argument from stack. This is a MemberName which we resolve to the
+  // target method.
+  oop vmentry = popFromStack(THREAD);
+
+  Method* vmtarget = (Method*) java_lang_invoke_MemberName::vmtarget(vmentry);
+
+  invoke_target(vmtarget, THREAD);
+
+  return 0;
 }
 
-void MethodHandles::generate_method_handle_stub(MacroAssembler*          masm,
-                                                MethodHandles::EntryKind ek) {
-  init_entry(ek, (MethodHandleEntry *) ek);
+int MethodHandles::method_handle_entry_linkToInterface(Method* method, intptr_t UNUSED, TRAPS) {
+  JavaThread *thread = (JavaThread *) THREAD;
+  InterpreterFrame *frame = thread->top_zero_frame()->as_interpreter_frame();
+  interpreterState istate = frame->interpreter_state();
+
+  // Pop appendix argument from stack. This is a MemberName which we resolve to the
+  // target method.
+  oop vmentry = popFromStack(THREAD);
+  intptr_t* topOfStack = istate->stack();
+
+  // Resolve target method by looking up in the receiver object's itable.
+  Klass* clazz = java_lang_Class::as_Klass(java_lang_invoke_MemberName::clazz(vmentry));
+  intptr_t vmindex = java_lang_invoke_MemberName::vmindex(vmentry);
+  Method* target = (Method*) java_lang_invoke_MemberName::vmtarget(vmentry);
+
+  int numArgs = target->size_of_parameters();
+  oop recv = STACK_OBJECT(-numArgs);
+
+  InstanceKlass* klass_part = InstanceKlass::cast(recv->klass());
+  itableOffsetEntry* ki = (itableOffsetEntry*) klass_part->start_of_itable();
+  int i;
+  for ( i = 0 ; i < klass_part->itable_length() ; i++, ki++ ) {
+    if (ki->interface_klass() == clazz) break;
+  }
+
+  itableMethodEntry* im = ki->first_method_entry(recv->klass());
+  Method* vmtarget = im[vmindex].method();
+
+  invoke_target(vmtarget, THREAD);
+
+  return 0;
 }
+
+int MethodHandles::method_handle_entry_linkToVirtual(Method* method, intptr_t UNUSED, TRAPS) {
+  JavaThread *thread = (JavaThread *) THREAD;
+
+  InterpreterFrame *frame = thread->top_zero_frame()->as_interpreter_frame();
+  interpreterState istate = frame->interpreter_state();
+
+  // Pop appendix argument from stack. This is a MemberName which we resolve to the
+  // target method.
+  oop vmentry = popFromStack(THREAD);
+  intptr_t* topOfStack = istate->stack();
+
+  // Resolve target method by looking up in the receiver object's vtable.
+  intptr_t vmindex = java_lang_invoke_MemberName::vmindex(vmentry);
+  Method* target = (Method*) java_lang_invoke_MemberName::vmtarget(vmentry);
+  int numArgs = target->size_of_parameters();
+  oop recv = STACK_OBJECT(-numArgs);
+  Klass* clazz = recv->klass();
+  Klass* klass_part = InstanceKlass::cast(clazz);
+  klassVtable* vtable = klass_part->vtable();
+  Method* vmtarget = vtable->method_at(vmindex);
+
+  invoke_target(vmtarget, THREAD);
+
+  return 0;
+}
+
+int MethodHandles::method_handle_entry_invalid(Method* method, intptr_t UNUSED, TRAPS) {
+  ShouldNotReachHere();
+  return 0;
+}
+
+address MethodHandles::generate_method_handle_interpreter_entry(MacroAssembler* masm,
+                                                                vmIntrinsics::ID iid) {
+  switch (iid) {
+  case vmIntrinsics::_invokeGeneric:
+  case vmIntrinsics::_compiledLambdaForm:
+    // Perhaps surprisingly, the symbolic references visible to Java are not directly used.
+    // They are linked to Java-generated adapters via MethodHandleNatives.linkMethod.
+    // They all allow an appendix argument.
+    return InterpreterGenerator::generate_entry_impl(masm, (address) MethodHandles::method_handle_entry_invalid);
+  case vmIntrinsics::_invokeBasic:
+    return InterpreterGenerator::generate_entry_impl(masm, (address) MethodHandles::method_handle_entry_invokeBasic);
+  case vmIntrinsics::_linkToStatic:
+  case vmIntrinsics::_linkToSpecial:
+    return InterpreterGenerator::generate_entry_impl(masm, (address) MethodHandles::method_handle_entry_linkToStaticOrSpecial);
+  case vmIntrinsics::_linkToInterface:
+    return InterpreterGenerator::generate_entry_impl(masm, (address) MethodHandles::method_handle_entry_linkToInterface);
+  case vmIntrinsics::_linkToVirtual:
+    return InterpreterGenerator::generate_entry_impl(masm, (address) MethodHandles::method_handle_entry_linkToVirtual);
+  default:
+    ShouldNotReachHere();
+    return NULL;
+  }
+}

--- a/src/cpu/zero/vm/methodHandles_zero.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/methodHandles_zero.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -26,6 +26,14 @@
 
 // Adapters
 enum /* platform_dependent_constants */ {
-  adapter_code_size = 0
+  adapter_code_size = sizeof(ZeroEntry) * (Interpreter::method_handle_invoke_LAST - Interpreter::method_handle_invoke_FIRST + 1)
 };
 
+private:
+  static oop popFromStack(TRAPS);
+  static void invoke_target(Method* method, TRAPS);
+  static int method_handle_entry_invokeBasic(Method* method, intptr_t UNUSED, TRAPS);
+  static int method_handle_entry_linkToStaticOrSpecial(Method* method, intptr_t UNUSED, TRAPS);
+  static int method_handle_entry_linkToVirtual(Method* method, intptr_t UNUSED, TRAPS);
+  static int method_handle_entry_linkToInterface(Method* method, intptr_t UNUSED, TRAPS);
+  static int method_handle_entry_invalid(Method* method, intptr_t UNUSED, TRAPS);

--- a/src/cpu/zero/vm/register_zero.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/register_zero.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -114,5 +114,8 @@
 };
 
 CONSTANT_REGISTER_DECLARATION(Register, noreg, (-1));
+#ifndef DONT_USE_REGISTER_DEFINES
+#define noreg ((Register)(noreg_RegisterEnumValue))
+#endif
 
 #endif // CPU_ZERO_VM_REGISTER_ZERO_HPP

--- a/src/cpu/zero/vm/relocInfo_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/relocInfo_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -77,3 +77,7 @@
                                                        CodeBuffer*       dst) {
   ShouldNotCallThis();
 }
+
+void metadata_Relocation::pd_fix_value(address x) {
+  ShouldNotCallThis();
+}

--- a/src/cpu/zero/vm/sharedRuntime_zero.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/cpu/zero/vm/sharedRuntime_zero.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -35,6 +35,7 @@
 #include "runtime/sharedRuntime.hpp"
 #include "runtime/vframeArray.hpp"
 #include "vmreg_zero.inline.hpp"
+
 #ifdef COMPILER1
 #include "c1/c1_Runtime1.hpp"
 #endif
@@ -47,6 +48,12 @@
 #endif
 
 
+
+static address zero_null_code_stub() {
+  address start = ShouldNotCallThisStub();
+  return start;
+}
+
 int SharedRuntime::java_calling_convention(const BasicType *sig_bt,
                                            VMRegPair *regs,
                                            int total_args_passed,
@@ -63,16 +70,14 @@
                         AdapterFingerPrint *fingerprint) {
   return AdapterHandlerLibrary::new_entry(
     fingerprint,
-    ShouldNotCallThisStub(),
-    ShouldNotCallThisStub(),
-    ShouldNotCallThisStub());
+    CAST_FROM_FN_PTR(address,zero_null_code_stub),
+    CAST_FROM_FN_PTR(address,zero_null_code_stub),
+    CAST_FROM_FN_PTR(address,zero_null_code_stub));
 }
 
 nmethod *SharedRuntime::generate_native_wrapper(MacroAssembler *masm,
                                                 methodHandle method,
                                                 int compile_id,
-                                                int total_args_passed,
-                                                int max_arg,
                                                 BasicType *sig_bt,
                                                 VMRegPair *regs,
                                                 BasicType ret_type) {
@@ -96,19 +101,20 @@
   ShouldNotCallThis();
 }
 
+JRT_LEAF(void, zero_stub())
+  ShouldNotCallThis();
+JRT_END
+
 static RuntimeStub* generate_empty_runtime_stub(const char* name) {
-  CodeBuffer buffer(name, 0, 0);
-  return RuntimeStub::new_runtime_stub(name, &buffer, 0, 0, NULL, false);
+  return CAST_FROM_FN_PTR(RuntimeStub*,zero_stub);
 }
 
 static SafepointBlob* generate_empty_safepoint_blob() {
-  CodeBuffer buffer("handler_blob", 0, 0);
-  return SafepointBlob::create(&buffer, NULL, 0);
+  return CAST_FROM_FN_PTR(SafepointBlob*,zero_stub);
 }
 
 static DeoptimizationBlob* generate_empty_deopt_blob() {
-  CodeBuffer buffer("handler_blob", 0, 0);
-  return DeoptimizationBlob::create(&buffer, NULL, 0, 0, 0, 0);
+  return CAST_FROM_FN_PTR(DeoptimizationBlob*,zero_stub);
 }
 
 
@@ -116,7 +122,7 @@
   _deopt_blob = generate_empty_deopt_blob();
 }
 
-SafepointBlob* SharedRuntime::generate_handler_blob(address call_ptr, bool cause_return) {
+SafepointBlob* SharedRuntime::generate_handler_blob(address call_ptr, int poll_type) {
   return generate_empty_safepoint_blob();
 }
 
@@ -124,6 +130,7 @@
   return generate_empty_runtime_stub("resolve_blob");
 }
 
+
 int SharedRuntime::c_calling_convention(const BasicType *sig_bt,
                                          VMRegPair *regs,
                                          int total_args_passed) {

--- a/src/os/bsd/vm/attachListener_bsd.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/os/bsd/vm/attachListener_bsd.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -342,7 +342,6 @@
 
     // get the credentials of the peer and check the effective uid/guid
     // - check with jeff on this.
-#ifdef _ALLBSD_SOURCE
     uid_t puid;
     gid_t pgid;
     if (::getpeereid(s, &puid, &pgid) != 0) {
@@ -350,17 +349,6 @@
       RESTARTABLE(::close(s), res);
       continue;
     }
-#else
-    struct ucred cred_info;
-    socklen_t optlen = sizeof(cred_info);
-    if (::getsockopt(s, SOL_SOCKET, SO_PEERCRED, (void*)&cred_info, &optlen) == -1) {
-      int res;
-      RESTARTABLE(::close(s), res);
-      continue;
-    }
-    uid_t puid = cred_info.uid;
-    gid_t pgid = cred_info.gid;
-#endif
     uid_t euid = geteuid();
     gid_t egid = getegid();
 

--- a/src/os/bsd/vm/osThread_bsd.hpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/os/bsd/vm/osThread_bsd.hpp	Mon Nov 05 13:55:31 2012 -0800
@@ -39,18 +39,12 @@
 
  private:
 
-#ifdef _ALLBSD_SOURCE
-
 #ifdef __APPLE__
   typedef thread_t thread_id_t;
 #else
   typedef pthread_t thread_id_t;
 #endif
 
-#else
-  typedef pid_t thread_id_t;
-#endif
-
   // _pthread_id is the pthread id, which is used by library calls
   // (e.g. pthread_kill).
   pthread_t _pthread_id;

--- a/src/os/bsd/vm/os_bsd.cpp	Mon Nov 05 15:30:22 2012 -0500
+++ b/src/os/bsd/vm/os_bsd.cpp	Mon Nov 05 13:55:31 2012 -0800
@@ -108,14 +108,8 @@
 # include <semaphore.h>
 # include <fcntl.h>
 # include <string.h>
-#ifdef _ALLBSD_SOURCE
 # include <sys/param.h>
 # include <sys/sysctl.h>
-#else
-# include <syscall.h>
-# include <sys/sysinfo.h>
-# include <gnu/libc-version.h>
-#endif
 # include <sys/ipc.h>
 # include <sys/shm.h>
 #ifndef __APPLE__
@@ -150,25 +144,10 @@
 // global variables
 julong os::Bsd::_physical_memory = 0;
 
-#ifndef _ALLBSD_SOURCE
-address   os::Bsd::_initial_thread_stack_bottom = NULL;
-uintptr_t os::Bsd::_initial_thread_stack_size   = 0;
-#endif
 
 int (*os::Bsd::_clock_gettime)(clockid_t, struct timespec *) = NULL;
-#ifndef _ALLBSD_SOURCE
-int (*os::Bsd::_pthread_getcpuclockid)(pthread_t, clockid_t *) = NULL;
-Mutex* os::Bsd::_createThread_lock = NULL;
-#endif
 pthread_t os::Bsd::_main_thread;
 int os::Bsd::_page_size = -1;
-#ifndef _ALLBSD_SOURCE
-bool os::Bsd::_is_floating_stack = false;
-bool os::Bsd::_is_NPTL = false;
-bool os::Bsd::_supports_fast_thread_cpu_time = false;
-const char * os::Bsd::_glibc_version = NULL;
-const char * os::Bsd::_libpthread_version = NULL;
-#endif
 
 static jlong initial_time_count=0;
 
@@ -176,7 +155,7 @@
 
 // For diagnostics to print a message once. see run_periodic_checks
 static sigset_t check_signal_done;
-static bool check_signals = true;;
+static bool check_signals = true;
 
 static pid_t _initial_pid = 0;
 
@@ -198,16 +177,8 @@
 }
 
 julong os::Bsd::available_memory() {
-#ifdef _ALLBSD_SOURCE
   // XXXBSD: this is just a stopgap implementation
   return physical_memory() >> 2;
-#else
-  // values in struct sysinfo are "unsigned long"
-  struct sysinfo si;
-  sysinfo(&si);
-
-  return (julong)si.freeram * si.mem_unit;
-#endif
 }
 
 julong os::physical_memory() {
@@ -255,22 +226,6 @@
 }
 
 
-#ifndef _ALLBSD_SOURCE
-#ifndef SYS_gettid
-// i386: 224, ia64: 1105, amd64: 186, sparc 143
-#ifdef __ia64__
-#define SYS_gettid 1105
-#elif __i386__
-#define SYS_gettid 224
-#elif __amd64__
-#define SYS_gettid 186
-#elif __sparc__
-#define SYS_gettid 143
-#else
-#error define gettid for the arch
-#endif
-#endif
-#endif
 
 // Cpu architecture string
 #if   defined(ZERO)
@@ -302,36 +257,7 @@
 #define COMPILER_VARIANT "client"
 #endif
 
-#ifndef _ALLBSD_SOURCE
-// pid_t gettid()
-//
-// Returns the kernel thread id of the currently running thread. Kernel
-// thread id is used to access /proc.
-//
-// (Note that getpid() on BsdThreads returns kernel thread id too; but
-// on NPTL, it returns the same pid for all threads, as required by POSIX.)
-//
-pid_t os::Bsd::gettid() {
-  int rslt = syscall(SYS_gettid);
-  if (rslt == -1) {
-     // old kernel, no NPTL support
-     return getpid();
-  } else {
-     return (pid_t)rslt;
-  }
-}
-
-// Most versions of bsd have a bug where the number of processors are
-// determined by looking at the /proc file system.  In a chroot environment,
-// the system call returns 1.  This causes the VM to act as if it is
-// a single processor and elide locking (see is_MP() call).
-static bool unsafe_chroot_detected = false;
-static const char *unstable_chroot_error = "/proc file system not found.\n"
-                     "Java may be unstable running multithreaded in a chroot "
-                     "environment on Bsd when /proc filesystem is not mounted.";
-#endif
-
-#ifdef _ALLBSD_SOURCE
+
 void os::Bsd::initialize_system_info() {
   int mib[2];
   size_t len;
@@ -370,24 +296,6 @@
   }
 #endif
 }
-#else
-void os::Bsd::initialize_system_info() {
-  set_processor_count(sysconf(_SC_NPROCESSORS_CONF));
-  if (processor_count() == 1) {
-    pid_t pid = os::Bsd::gettid();
-    char fname[32];
-    jio_snprintf(fname, sizeof(fname), "/proc/%d", pid);
-    FILE *fp = fopen(fname, "r");
-    if (fp == NULL) {
-      unsafe_chroot_detected = true;
-    } else {
-      fclose(fp);
-    }
-  }
-  _physical_memory = (julong)sysconf(_SC_PHYS_PAGES) * (julong)sysconf(_SC_PAGESIZE);
-  assert(processor_count() > 0, "bsd error");
-}
-#endif
 
 #ifdef __APPLE__
 static const char *get_home() {
@@ -744,171 +652,6 @@
   }
 }
 
-#ifndef _ALLBSD_SOURCE
-//////////////////////////////////////////////////////////////////////////////
-// detecting pthread library
-
-void os::Bsd::libpthread_init() {
-  // Save glibc and pthread version strings. Note that _CS_GNU_LIBC_VERSION
-  // and _CS_GNU_LIBPTHREAD_VERSION are supported in glibc >= 2.3.2. Use a
-  // generic name for earlier versions.
-  // Define macros here so we can build HotSpot on old systems.
-# ifndef _CS_GNU_LIBC_VERSION
-# define _CS_GNU_LIBC_VERSION 2
-# endif
-# ifndef _CS_GNU_LIBPTHREAD_VERSION
-# define _CS_GNU_LIBPTHREAD_VERSION 3
-# endif
-
-  size_t n = confstr(_CS_GNU_LIBC_VERSION, NULL, 0);
-  if (n > 0) {
-     char *str = (char *)malloc(n);
-     confstr(_CS_GNU_LIBC_VERSION, str, n);
-     os::Bsd::set_glibc_version(str);
-  } else {
-     // _CS_GNU_LIBC_VERSION is not supported, try gnu_get_libc_version()
-     static char _gnu_libc_version[32];
-     jio_snprintf(_gnu_libc_version, sizeof(_gnu_libc_version),
-              "glibc %s %s", gnu_get_libc_version(), gnu_get_libc_release());
-     os::Bsd::set_glibc_version(_gnu_libc_version);
-  }
-
-  n = confstr(_CS_GNU_LIBPTHREAD_VERSION, NULL, 0);
-  if (n > 0) {
-     char *str = (char *)malloc(n);
-     confstr(_CS_GNU_LIBPTHREAD_VERSION, str, n);
-     // Vanilla RH-9 (glibc 2.3.2) has a bug that confstr() always tells
-     // us "NPTL-0.29" even we are running with BsdThreads. Check if this
-     // is the case. BsdThreads has a hard limit on max number of threads.
-     // So sysconf(_SC_THREAD_THREADS_MAX) will return a positive value.
-     // On the other hand, NPTL does not have such a limit, sysconf()
-     // will return -1 and errno is not changed. Check if it is really NPTL.
-     if (strcmp(os::Bsd::glibc_version(), "glibc 2.3.2") == 0 &&
-         strstr(str, "NPTL") &&
-         sysconf(_SC_THREAD_THREADS_MAX) > 0) {
-       free(str);
-       os::Bsd::set_libpthread_version("bsdthreads");
-     } else {
-       os::Bsd::set_libpthread_version(str);
-     }
-  } else {
-    // glibc before 2.3.2 only has BsdThreads.
-    os::Bsd::set_libpthread_version("bsdthreads");
-  }
-
-  if (strstr(libpthread_version(), "NPTL")) {
-     os::Bsd::set_is_NPTL();
-  } else {
-     os::Bsd::set_is_BsdThreads();
-  }
-
-  // BsdThreads have two flavors: floating-stack mode, which allows variable
-  // stack size; and fixed-stack mode. NPTL is always floating-stack.
-  if (os::Bsd::is_NPTL() || os::Bsd::supports_variable_stack_size()) {
-     os::Bsd::set_is_floating_stack();
-  }
-}
-
-/////////////////////////////////////////////////////////////////////////////
-// thread stack
-
-// Force Bsd kernel to expand current thread stack. If "bottom" is close
-// to the stack guard, caller should block all signals.
-//
-// MAP_GROWSDOWN:
-//   A special mmap() flag that is used to implement thread stacks. It tells
-//   kernel that the memory region should extend downwards when needed. This
-//   allows early versions of BsdThreads to only mmap the first few pages
-//   when creating a new thread. Bsd kernel will automatically expand thread
-//   stack as needed (on page faults).
-//
-//   However, because the memory region of a MAP_GROWSDOWN stack can grow on
-//   demand, if a page fault happens outside an already mapped MAP_GROWSDOWN
-//   region, it's hard to tell if the fault is due to a legitimate stack
-//   access or because of reading/writing non-exist memory (e.g. buffer
-//   overrun). As a rule, if the fault happens below current stack pointer,
-//   Bsd kernel does not expand stack, instead a SIGSEGV is sent to the
-//   application (see Bsd kernel fault.c).
-//
-//   This Bsd feature can cause SIGSEGV when VM bangs thread stack for
-//   stack overflow detection.
-//
-//   Newer version of BsdThreads (since glibc-2.2, or, RH-7.x) and NPTL do
-//   not use this flag. However, the stack of initial thread is not created
-//   by pthread, it is still MAP_GROWSDOWN. Also it's possible (though
-//   unlikely) that user code can create a thread with MAP_GROWSDOWN stack
-//   and then attach the thread to JVM.
-//
-// To get around the problem and allow stack banging on Bsd, we need to
-// manually expand thread stack after receiving the SIGSEGV.
-//
-// There are two ways to expand thread stack to address "bottom", we used
-// both of them in JVM before 1.5:
-//   1. adjust stack pointer first so that it is below "bottom", and then
-//      touch "bottom"
-//   2. mmap() the page in question
-//
-// Now alternate signal stack is gone, it's harder to use 2. For instance,
-// if current sp is already near the lower end of page 101, and we need to
-// call mmap() to map page 100, it is possible that part of the mmap() frame
-// will be placed in page 100. When page 100 is mapped, it is zero-filled.
-// That will destroy the mmap() frame and cause VM to crash.
-//
-// The following code works by adjusting sp first, then accessing the "bottom"
-// page to force a page fault. Bsd kernel will then automatically expand the
-// stack mapping.
-//
-// _expand_stack_to() assumes its frame size is less than page size, which
-// should always be true if the function is not inlined.
-
-#if __GNUC__ < 3    // gcc 2.x does not support noinline attribute
-#define NOINLINE
-#else
-#define NOINLINE __attribute__ ((noinline))
-#endif
-
-static void _expand_stack_to(address bottom) NOINLINE;
-
-static void _expand_stack_to(address bottom) {
-  address sp;
-  size_t size;
-  volatile char *p;
-
-  // Adjust bottom to point to the largest address within the same page, it
-  // gives us a one-page buffer if alloca() allocates slightly more memory.
-  bottom = (address)align_size_down((uintptr_t)bottom, os::Bsd::page_size());
-  bottom += os::Bsd::page_size() - 1;
-
-  // sp might be slightly above current stack pointer; if that's the case, we
-  // will alloca() a little more space than necessary, which is OK. Don't use
-  // os::current_stack_pointer(), as its result can be slightly below current
-  // stack pointer, causing us to not alloca enough to reach "bottom".
-  sp = (address)&sp;
-
-  if (sp > bottom) {
-    size = sp - bottom;
-    p = (volatile char *)alloca(size);
-    assert(p != NULL && p <= (volatile char *)bottom, "alloca problem?");
-    p[0] = '\0';
-  }
-}
-
-bool os::Bsd::manually_expand_stack(JavaThread * t, address addr) {
-  assert(t!=NULL, "just checking");
-  assert(t->osthread()->expanding_stack(), "expand should be set");
-  assert(t->stack_base() != NULL, "stack_base was not initialized");
-
-  if (addr <  t->stack_base() && addr >= t->stack_yellow_zone_base()) {
-    sigset_t mask_all, old_sigset;
-    sigfillset(&mask_all);
-    pthread_sigmask(SIG_SETMASK, &mask_all, &old_sigset);
-    _expand_stack_to(addr);
-    pthread_sigmask(SIG_SETMASK, &old_sigset, NULL);
-    return true;
-  }
-  return false;
-}
-#endif
 
 //////////////////////////////////////////////////////////////////////////////
 // create new thread
@@ -917,43 +660,7 @@
 
 // check if it's safe to start a new thread
 static bool _thread_safety_check(Thread* thread) {
-#ifdef _ALLBSD_SOURCE
-    return true;
-#else
-  if (os::Bsd::is_BsdThreads() && !os::Bsd::is_floating_stack()) {
-    // Fixed stack BsdThreads (SuSE Bsd/x86, and some versions of Redhat)
-    //   Heap is mmap'ed at lower end of memory space. Thread stacks are
-    //   allocated (MAP_FIXED) from high address space. Every thread stack
-    //   occupies a fixed size slot (usually 2Mbytes, but user can change
-    //   it to other values if they rebuild BsdThreads).
-    //
-    // Problem with MAP_FIXED is that mmap() can still succeed even part of
-    // the memory region has already been mmap'ed. That means if we have too
-    // many threads and/or very large heap, eventually thread stack will
-    // collide with heap.
-    //
-    // Here we try to prevent heap/stack collision by comparing current
-    // stack bottom with the highest address that has been mmap'ed by JVM
-    // plus a safety margin for memory maps created by native code.
-    //
-    // This feature can be disabled by setting ThreadSafetyMargin to 0
-    //
-    if (ThreadSafetyMargin > 0) {
-      address stack_bottom = os::current_stack_base() - os::current_stack_size();
-
-      // not safe if our stack extends below the safety margin
-      return stack_bottom - ThreadSafetyMargin >= highest_vm_reserved_address();
-    } else {
-      return true;
-    }
-  } else {
-    // Floating stack BsdThreads or NPTL:
-    //   Unlike fixed stack BsdThreads, thread stacks are not MAP_FIXED. When
-    //   there's not enough space left, pthread_create() will fail. If we come
-    //   here, that means enough space has been reserved for stack.
-    return true;
-  }
-#endif
+  return true;
 }
 
 #ifdef __APPLE__
@@ -991,7 +698,6 @@
     return NULL;
   }
 
-#ifdef _ALLBSD_SOURCE
 #ifdef __APPLE__
   // thread_id is mach thread on macos
   osthread->set_thread_id(::mach_thread_self());
@@ -999,17 +705,6 @@
   // thread_id is pthread_id on BSD
   osthread->set_thread_id(::pthread_self());
 #endif
-#else
-  // thread_id is kernel thread id (similar to Solaris LWP id)
-  osthread->set_thread_id(os::Bsd::gettid());
-
-  if (UseNUMA) {
-    int lgrp_id = os::numa_get_group_id();
-    if (lgrp_id != -1) {
-      thread->set_lgrp_id(lgrp_id);
-    }
-  }
-#endif
   // initialize signal mask for this thread
   os::Bsd::hotspot_sigmask(thread);
 
@@ -1099,23 +794,9 @@
     // let pthread_create() pick the default value.
   }
 
-#ifndef _ALLBSD_SOURCE
-  // glibc guard page
-  pthread_attr_setguardsize(&attr, os::Bsd::default_guard_size(thr_type));
-#endif
-
   ThreadState state;
 
   {
-
-#ifndef _ALLBSD_SOURCE
-    // Serialize thread creation if we are running with fixed stack BsdThreads
-    bool lock = os::Bsd::is_BsdThreads() && !os::Bsd::is_floating_stack();
-    if (lock) {
-      os::Bsd::createThread_lock()->lock_without_safepoint_check();
-    }
-#endif
-
     pthread_t tid;
     int ret = pthread_create(&tid, &attr, (void* (*)(void*)) java_start, thread);
 
@@ -1128,9 +809,6 @@
       // Need to clean up stuff we've allocated so far
       thread->set_osthread(NULL);
       delete osthread;
-#ifndef _ALLBSD_SOURCE
-      if (lock) os::Bsd::createThread_lock()->unlock();
-#endif
       return false;
     }
 
@@ -1146,11 +824,6 @@
       }
     }
 
-#ifndef _ALLBSD_SOURCE
-    if (lock) {
-      os::Bsd::createThread_lock()->unlock();
-    }
-#endif
   }
 
   // Aborted due to thread limit being reached
@@ -1188,15 +861,11 @@
   }
 
   // Store pthread info into the OSThread
-#ifdef _ALLBSD_SOURCE
 #ifdef __APPLE__
   osthread->set_thread_id(::mach_thread_self());
 #else
   osthread->set_thread_id(::pthread_self());
 #endif
-#else
-  osthread->set_thread_id(os::Bsd::gettid());
-#endif
   osthread->set_pthread_id(::pthread_self());
 
   // initialize floating point control register
@@ -1207,35 +876,6 @@
 
   thread->set_osthread(osthread);
 
-#ifndef _ALLBSD_SOURCE
-  if (UseNUMA) {
-    int lgrp_id = os::numa_get_group_id();
-    if (lgrp_id != -1) {
-      thread->set_lgrp_id(lgrp_id);
-    }
-  }
-
-  if (os::Bsd::is_initial_thread()) {
-    // If current thread is initial thread, its stack is mapped on demand,
-    // see notes about MAP_GROWSDOWN. Here we try to force kernel to map
-    // the entire stack region to avoid SEGV in stack banging.
-    // It is also useful to get around the heap-stack-gap problem on SuSE
-    // kernel (see 4821821 for details). We first expand stack to the top
-    // of yellow zone, then enable stack yellow zone (order is significant,
-    // enabling yellow zone first will crash JVM on SuSE Bsd), so there
-    // is no gap between the last two virtual memory regions.
-
-    JavaThread *jt = (JavaThread *)thread;
-    address addr = jt->stack_yellow_zone_base();
-    assert(addr != NULL, "initialization problem?");
-    assert(jt->stack_available(addr) > 0, "stack guard should not be enabled");
-
-    osthread->set_expanding_stack();
-    os::Bsd::manually_expand_stack(jt, addr);
-    osthread->clear_expanding_stack();
-  }
-#endif
-
   // initialize signal mask for this thread
   // and save the caller's signal mask
   os::Bsd::hotspot_sigmask(thread);
@@ -1290,247 +930,6 @@
   return ThreadLocalStorage::thread();
 }
 
-//////////////////////////////////////////////////////////////////////////////
-// initial thread
-
-#ifndef _ALLBSD_SOURCE
-// Check if current thread is the initial thread, similar to Solaris thr_main.
-bool os::Bsd::is_initial_thread(void) {
-  char dummy;
-  // If called before init complete, thread stack bottom will be null.
-  // Can be called if fatal error occurs before initialization.
-  if (initial_thread_stack_bottom() == NULL) return false;
-  assert(initial_thread_stack_bottom() != NULL &&
-         initial_thread_stack_size()   != 0,
-         "os::init did not locate initial thread's stack region");
-  if ((address)&dummy >= initial_thread_stack_bottom() &&
-      (address)&dummy < initial_thread_stack_bottom() + initial_thread_stack_size())
-       return true;
-  else return false;
-}
-
-// Find the virtual memory area that contains addr
-static bool find_vma(address addr, address* vma_low, address* vma_high) {
-  FILE *fp = fopen("/proc/self/maps", "r");
-  if (fp) {
-    address low, high;
-    while (!feof(fp)) {
-      if (fscanf(fp, "%p-%p", &low, &high) == 2) {
-        if (low <= addr && addr < high) {
-           if (vma_low)  *vma_low  = low;
-           if (vma_high) *vma_high = high;
-           fclose (fp);
-           return true;
-        }
-      }
-      for (;;) {
-        int ch = fgetc(fp);
-        if (ch == EOF || ch == (int)'\n') break;
-      }
-    }
-    fclose(fp);
-  }
-  return false;
-}
-
-// Locate initial thread stack. This special handling of initial thread stack
-// is needed because pthread_getattr_np() on most (all?) Bsd distros returns
-// bogus value for initial thread.
-void os::Bsd::capture_initial_stack(size_t max_size) {
-  // stack size is the easy part, get it from RLIMIT_STACK
-  size_t stack_size;
-  struct rlimit rlim;
-  getrlimit(RLIMIT_STACK, &rlim);
-  stack_size = rlim.rlim_cur;
-
-  // 6308388: a bug in ld.so will relocate its own .data section to the
-  //   lower end of primordial stack; reduce ulimit -s value a little bit
-  //   so we won't install guard page on ld.so's data section.
-  stack_size -= 2 * page_size();
-
-  // 4441425: avoid crash with "unlimited" stack size on SuSE 7.1 or Redhat
-  //   7.1, in both cases we will get 2G in return value.
-  // 4466587: glibc 2.2.x compiled w/o "--enable-kernel=2.4.0" (RH 7.0,
-  //   SuSE 7.2, Debian) can not handle alternate signal stack correctly
-  //   for initial thread if its stack size exceeds 6M. Cap it at 2M,
-  //   in case other parts in glibc still assumes 2M max stack size.
-  // FIXME: alt signal stack is gone, maybe we can relax this constraint?
-#ifndef IA64
-  if (stack_size > 2 * K * K) stack_size = 2 * K * K;
-#else
-  // Problem still exists RH7.2 (IA64 anyway) but 2MB is a little small
-  if (stack_size > 4 * K * K) stack_size = 4 * K * K;
-#endif
-
-  // Try to figure out where the stack base (top) is. This is harder.
-  //
-  // When an application is started, glibc saves the initial stack pointer in
-  // a global variable "__libc_stack_end", which is then used by system
-  // libraries. __libc_stack_end should be pretty close to stack top. The
-  // variable is available since the very early days. However, because it is
-  // a private interface, it could disappear in the future.
-  //
-  // Bsd kernel saves start_stack information in /proc/<pid>/stat. Similar
-  // to __libc_stack_end, it is very close to stack top, but isn't the real
-  // stack top. Note that /proc may not exist if VM is running as a chroot
-  // program, so reading /proc/<pid>/stat could fail. Also the contents of
-  // /proc/<pid>/stat could change in the future (though unlikely).
-  //
-  // We try __libc_stack_end first. If that doesn't work, look for
-  // /proc/<pid>/stat. If neither of them works, we use current stack pointer
-  // as a hint, which should work well in most cases.
-
-  uintptr_t stack_start;
-
-  // try __libc_stack_end first
-  uintptr_t *p = (uintptr_t *)dlsym(RTLD_DEFAULT, "__libc_stack_end");
-  if (p && *p) {
-    stack_start = *p;
-  } else {
-    // see if we can get the start_stack field from /proc/self/stat
-    FILE *fp;
-    int pid;
-    char state;
-    int ppid;
-    int pgrp;
-    int session;
-    int nr;
-    int tpgrp;
-    unsigned long flags;
-    unsigned long minflt;
-    unsigned long cminflt;
-    unsigned long majflt;
-    unsigned long cmajflt;
-    unsigned long utime;
-    unsigned long stime;
-    long cutime;
-    long cstime;
-    long prio;
-    long nice;
-    long junk;
-    long it_real;
-    uintptr_t start;
-    uintptr_t vsize;
-    intptr_t rss;
-    uintptr_t rsslim;
-    uintptr_t scodes;
-    uintptr_t ecode;
-    int i;
-
-    // Figure what the primordial thread stack base is. Code is inspired
-    // by email from Hans Boehm. /proc/self/stat begins with current pid,
-    // followed by command name surrounded by parentheses, state, etc.
-    char stat[2048];
-    int statlen;
-
-    fp = fopen("/proc/self/stat", "r");
-    if (fp) {
-      statlen = fread(stat, 1, 2047, fp);
-      stat[statlen] = '\0';
-      fclose(fp);
-
-      // Skip pid and the command string. Note that we could be dealing with
-      // weird command names, e.g. user could decide to rename java launcher
-      // to "java 1.4.2 :)", then the stat file would look like
-      //                1234 (java 1.4.2 :)) R ... ...
-      // We don't really need to know the command string, just find the last
-      // occurrence of ")" and then start parsing from there. See bug 4726580.
-      char * s = strrchr(stat, ')');
-
-      i = 0;
-      if (s) {
-        // Skip blank chars
-        do s++; while (isspace(*s));
-
-#define _UFM UINTX_FORMAT
-#define _DFM INTX_FORMAT
-
-        /*                                     1   1   1   1   1   1   1   1   1   1   2   2    2    2    2    2    2    2    2 */
-        /*              3  4  5  6  7  8   9   0   1   2   3   4   5   6   7   8   9   0   1    2    3    4    5    6    7    8 */
-        i = sscanf(s, "%c %d %d %d %d %d %lu %lu %lu %lu %lu %lu %lu %ld %ld %ld %ld %ld %ld " _UFM _UFM _DFM _UFM _UFM _UFM _UFM,
-             &state,          /* 3  %c  */
-             &ppid,           /* 4  %d  */
-             &pgrp,           /* 5  %d  */
-             &session,        /* 6  %d  */
-             &nr,             /* 7  %d  */
-             &tpgrp,          /* 8  %d  */
-             &flags,          /* 9  %lu  */
-             &minflt,         /* 10 %lu  */
-             &cminflt,        /* 11 %lu  */
-             &majflt,         /* 12 %lu  */
-             &cmajflt,        /* 13 %lu  */
-             &utime,          /* 14 %lu  */
-             &stime,          /* 15 %lu  */
-             &cutime,         /* 16 %ld  */
-             &cstime,         /* 17 %ld  */
-             &prio,           /* 18 %ld  */
-             &nice,           /* 19 %ld  */
-             &junk,           /* 20 %ld  */
-             &it_real,        /* 21 %ld  */
-             &start,          /* 22 UINTX_FORMAT */
-             &vsize,          /* 23 UINTX_FORMAT */
-             &rss,            /* 24 INTX_FORMAT  */
-             &rsslim,         /* 25 UINTX_FORMAT */
-             &scodes,         /* 26 UINTX_FORMAT */
-             &ecode,          /* 27 UINTX_FORMAT */
-             &stack_start);   /* 28 UINTX_FORMAT */
-      }
-
-#undef _UFM
-#undef _DFM
-
-      if (i != 28 - 2) {
-         assert(false, "Bad conversion from /proc/self/stat");
-         // product mode - assume we are the initial thread, good luck in the
-         // embedded case.
-         warning("Can't detect initial thread stack location - bad conversion");
-         stack_start = (uintptr_t) &rlim;
-      }
-    } else {
-      // For some reason we can't open /proc/self/stat (for example, running on
-      // FreeBSD with a Bsd emulator, or inside chroot), this should work for
-      // most cases, so don't abort:
-      warning("Can't detect initial thread stack location - no /proc/self/stat");
-      stack_start = (uintptr_t) &rlim;
-    }
-  }
-
-  // Now we have a pointer (stack_start) very close to the stack top, the
-  // next thing to do is to figure out the exact location of stack top. We
-  // can find out the virtual memory area that contains stack_start by
-  // reading /proc/self/maps, it should be the last vma in /proc/self/maps,
-  // and its upper limit is the real stack top. (again, this would fail if
-  // running inside chroot, because /proc may not exist.)
-
-  uintptr_t stack_top;
-  address low, high;
-  if (find_vma((address)stack_start, &low, &high)) {
-    // success, "high" is the true stack top. (ignore "low", because initial
-    // thread stack grows on demand, its real bottom is high - RLIMIT_STACK.)
-    stack_top = (uintptr_t)high;
-  } else {
-    // failed, likely because /proc/self/maps does not exist
-    warning("Can't detect initial thread stack location - find_vma failed");
-    // best effort: stack_start is normally within a few pages below the real
-    // stack top, use it as stack top, and reduce stack size so we won't put
-    // guard page outside stack.
-    stack_top = stack_start;
-    stack_size -= 16 * page_size();
-  }
-
-  // stack_top could be partially down the page so align it
-  stack_top = align_size_up(stack_top, page_size());
-
-  if (max_size && stack_size > max_size) {
-     _initial_thread_stack_size = max_size;
-  } else {
-     _initial_thread_stack_size = stack_size;
-  }
-
-  _initial_thread_stack_size = align_size_down(_initial_thread_stack_size, page_size());
-  _initial_thread_stack_bottom = (address)stack_top - _initial_thread_stack_size;
-}
-#endif
 
 ////////////////////////////////////////////////////////////////////////////////
 // time support
@@ -1576,7 +975,7 @@
 void os::Bsd::clock_init() {
         // XXXDARWIN: Investigate replacement monotonic clock
 }
-#elif defined(_ALLBSD_SOURCE)
+#else
 void os::Bsd::clock_init() {
   struct timespec res;
   struct timespec tp;
@@ -1586,86 +985,8 @@
     _clock_gettime = ::clock_gettime;
   }
 }
-#else
-void os::Bsd::clock_init() {
-  // we do dlopen's in this particular order due to bug in bsd
-  // dynamical loader (see 6348968) leading to crash on exit
-  void* handle = dlopen("librt.so.1", RTLD_LAZY);
-  if (handle == NULL) {
-    handle = dlopen("librt.so", RTLD_LAZY);
-  }
-
-  if (handle) {
-    int (*clock_getres_func)(clockid_t, struct timespec*) =
-           (int(*)(clockid_t, struct timespec*))dlsym(handle, "clock_getres");
-    int (*clock_gettime_func)(clockid_t, struct timespec*) =
-           (int(*)(clockid_t, struct timespec*))dlsym(handle, "clock_gettime");
-    if (clock_getres_func && clock_gettime_func) {
-      // See if monotonic clock is supported by the kernel. Note that some
-      // early implementations simply return kernel jiffies (updated every
-      // 1/100 or 1/1000 second). It would be bad to use such a low res clock
-      // for nano time (though the monotonic property is still nice to have).
-      // It's fixed in newer kernels, however clock_getres() still returns
-      // 1/HZ. We check if clock_getres() works, but will ignore its reported
-      // resolution for now. Hopefully as people move to new kernels, this
-      // won't be a problem.
-      struct timespec res;
-      struct timespec tp;
-      if (clock_getres_func (CLOCK_MONOTONIC, &res) == 0 &&
-          clock_gettime_func(CLOCK_MONOTONIC, &tp)  == 0) {
-        // yes, monotonic clock is supported
-        _clock_gettime = clock_gettime_func;
-      } else {
-        // close librt if there is no monotonic clock
-        dlclose(handle);
-      }
-    }
-  }
-}
 #endif
 
-#ifndef _ALLBSD_SOURCE
-#ifndef SYS_clock_getres
-
-#if defined(IA32) || defined(AMD64)
-#define SYS_clock_getres IA32_ONLY(266)  AMD64_ONLY(229)
-#define sys_clock_getres(x,y)  ::syscall(SYS_clock_getres, x, y)
-#else
-#warning "SYS_clock_getres not defined for this platform, disabling fast_thread_cpu_time"
-#define sys_clock_getres(x,y)  -1
-#endif
-
-#else
-#define sys_clock_getres(x,y)  ::syscall(SYS_clock_getres, x, y)
-#endif
-
-void os::Bsd::fast_thread_clock_init() {
-  if (!UseBsdPosixThreadCPUClocks) {
-    return;
-  }
-  clockid_t clockid;
-  struct timespec tp;
-  int (*pthread_getcpuclockid_func)(pthread_t, clockid_t *) =
-      (int(*)(pthread_t, clockid_t *)) dlsym(RTLD_DEFAULT, "pthread_getcpuclockid");
-
-  // Switch to using fast clocks for thread cpu time if
-  // the sys_clock_getres() returns 0 error code.
-  // Note, that some kernels may support the current thread
-  // clock (CLOCK_THREAD_CPUTIME_ID) but not the clocks
-  // returned by the pthread_getcpuclockid().
-  // If the fast Posix clocks are supported then the sys_clock_getres()
-  // must return at least tp.tv_sec == 0 which means a resolution
-  // better than 1 sec. This is extra check for reliability.
-
-  if(pthread_getcpuclockid_func &&
-     pthread_getcpuclockid_func(_main_thread, &clockid) == 0 &&
-     sys_clock_getres(clockid, &tp) == 0 && tp.tv_sec == 0) {
-
-    _supports_fast_thread_cpu_time = true;
-    _pthread_getcpuclockid = pthread_getcpuclockid_func;
-  }
-}
-#endif
 
 jlong os::javaTimeNanos() {
   if (Bsd::supports_monotonic_clock()) {
@@ -1978,7 +1299,6 @@
   return false;
 }
 
-#ifdef _ALLBSD_SOURCE
 // ported from solaris version
 bool os::dll_address_to_library_name(address addr, char* buf,
                                      int buflen, int* offset) {
@@ -1994,86 +1314,10 @@
      return false;
   }
 }
-#else
-struct _address_to_library_name {
-  address addr;          // input : memory address
-  size_t  buflen;        //         size of fname
-  char*   fname;         // output: library name
-  address base;          //         library base addr
-};
-
-static int address_to_library_name_callback(struct dl_phdr_info *info,
-                                            size_t size, void *data) {
-  int i;
-  bool found = false;
-  address libbase = NULL;
-  struct _address_to_library_name * d = (struct _address_to_library_name *)data;
-
-  // iterate through all loadable segments
-  for (i = 0; i < info->dlpi_phnum; i++) {
-    address segbase = (address)(info->dlpi_addr + info->dlpi_phdr[i].p_vaddr);
-    if (info->dlpi_phdr[i].p_type == PT_LOAD) {
-      // base address of a library is the lowest address of its loaded
-      // segments.
-      if (libbase == NULL || libbase > segbase) {
-        libbase = segbase;
-      }
-      // see if 'addr' is within current segment
-      if (segbase <= d->addr &&
-          d->addr < segbase + info->dlpi_phdr[i].p_memsz) {
-        found = true;
-      }
-    }
-  }
-
-  // dlpi_name is NULL or empty if the ELF file is executable, return 0
-  // so dll_address_to_library_name() can fall through to use dladdr() which
-  // can figure out executable name from argv[0].
-  if (found && info->dlpi_name && info->dlpi_name[0]) {
-    d->base = libbase;
-    if (d->fname) {
-      jio_snprintf(d->fname, d->buflen, "%s", info->dlpi_name);
-    }
-    return 1;
-  }
-  return 0;
-}
-
-bool os::dll_address_to_library_name(address addr, char* buf,
-                                     int buflen, int* offset) {
-  Dl_info dlinfo;
-  struct _address_to_library_name data;
-
-  // There is a bug in old glibc dladdr() implementation that it could resolve
-  // to wrong library name if the .so file has a base address != NULL. Here
-  // we iterate through the program headers of all loaded libraries to find
-  // out which library 'addr' really belongs to. This workaround can be
-  // removed once the minimum requirement for glibc is moved to 2.3.x.
-  data.addr = addr;
-  data.fname = buf;
-  data.buflen = buflen;
-  data.base = NULL;
-  int rslt = dl_iterate_phdr(address_to_library_name_callback, (void *)&data);
-
-  if (rslt) {
-     // buf already contains library name
-     if (offset) *offset = addr - data.base;
-     return true;
-  } else if (dladdr((void*)addr, &dlinfo)){
-     if (buf) jio_snprintf(buf, buflen, "%s", dlinfo.dli_fname);
-     if (offset) *offset = addr - (address)dlinfo.dli_fbase;
-     return true;
-  } else {
-     if (buf) buf[0] = '\0';
-     if (offset) *offset = -1;
-     return false;
-  }
-}
-#endif
-
-  // Loads .dll/.so and
-  // in case of error it checks if .dll/.so was built for the
-  // same architecture as Hotspot is running on
+
+// Loads .dll/.so and
+// in case of error it checks if .dll/.so was built for the
+// same architecture as Hotspot is running on
 
 #ifdef __APPLE__
 void * os::dll_load(const char *filename, char *ebuf, int ebuflen) {
@@ -2292,7 +1536,6 @@
 
 void os::print_dll_info(outputStream *st) {
    st->print_cr("Dynamic libraries:");
-#ifdef _ALLBSD_SOURCE
 #ifdef RTLD_DI_LINKMAP
     Dl_info dli;
     void *handle;
@@ -2336,16 +1579,6 @@
 #else
    st->print_cr("Error: Cannot print dynamic libraries.");
 #endif
-#else
-   char fname[32];
-   pid_t pid = os::Bsd::gettid();
-
-   jio_snprintf(fname, sizeof(fname), "/proc/%d/maps", pid);
-
-   if (!_print_ascii_file(fname, st)) {
-     st->print("Can not get library information for pid = %d\n", pid);
-   }
-#endif
 }
 
 void os::print_os_info_brief(outputStream* st) {
@@ -2374,22 +1607,10 @@
   st->print("Memory:");
   st->print(" %dk page", os::vm_page_size()>>10);
 
-#ifndef _ALLBSD_SOURCE
-  // values in struct sysinfo are "unsigned long"
-  struct sysinfo si;
-  sysinfo(&si);
-#endif
-
   st->print(", physical " UINT64_FORMAT "k",
             os::physical_memory() >> 10);
   st->print("(" UINT64_FORMAT "k free)",
             os::available_memory() >> 10);
-#ifndef _ALLBSD_SOURCE
-  st->print(", swap " UINT64_FORMAT "k",
-            ((jlong)si.totalswap * si.mem_unit) >> 10);
-  st->print("(" UINT64_FORMAT "k free)",
-            ((jlong)si.freeswap * si.mem_unit) >> 10);
-#endif
   st->cr();
 
   // meminfo
@@ -2786,42 +2007,13 @@
 #endif
 }
 
-#ifndef _ALLBSD_SOURCE
-// Define MAP_HUGETLB here so we can build HotSpot on old systems.
-#ifndef MAP_HUGETLB
-#define MAP_HUGETLB 0x40000
-#endif
-
-// Define MADV_HUGEPAGE here so we can build HotSpot on old systems.
-#ifndef MADV_HUGEPAGE
-#define MADV_HUGEPAGE 14
-#endif
-#endif
 
 bool os::pd_commit_memory(char* addr, size_t size, size_t alignment_hint,
                        bool exec) {
-#ifndef _ALLBSD_SOURCE
-  if (UseHugeTLBFS && alignment_hint > (size_t)vm_page_size()) {
-    int prot = exec ? PROT_READ|PROT_WRITE|PROT_EXEC : PROT_READ|PROT_WRITE;
-    uintptr_t res =
-      (uintptr_t) ::mmap(addr, size, prot,
-                         MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS|MAP_HUGETLB,
-                         -1, 0);
-    return res != (uintptr_t) MAP_FAILED;
-  }
-#endif
-
   return commit_memory(addr, size, exec);
 }
 
 void os::pd_realign_memory(char *addr, size_t bytes, size_t alignment_hint) {
-#ifndef _ALLBSD_SOURCE
-  if (UseHugeTLBFS && alignment_hint > (size_t)vm_page_size()) {
-    // We don't check the return value: madvise(MADV_HUGEPAGE) may not
-    // be supported or the memory may already be backed by huge pages.
-    ::madvise(addr, bytes, MADV_HUGEPAGE);
-  }
-#endif
 }
 
 void os::pd_free_memory(char *addr, size_t bytes, size_t alignment_hint) {
@@ -2860,111 +2052,6 @@
   return end;
 }
 
-#ifndef _ALLBSD_SOURCE
-// Something to do with the numa-aware allocator needs these symbols
-extern "C" JNIEXPORT void numa_warn(int number, char *where, ...) { }
-extern "C" JNIEXPORT void numa_error(char *where) { }
-extern "C" JNIEXPORT int fork1() { return fork(); }
-
-
-// If we are running with libnuma version > 2, then we should
-// be trying to use symbols with versions 1.1
-// If we are running with earlier version, which did not have symbol versions,
-// we should use the base version.
-void* os::Bsd::libnuma_dlsym(void* handle, const char *name) {
-  void *f = dlvsym(handle, name, "libnuma_1.1");
-  if (f == NULL) {
-    f = dlsym(handle, name);
-  }
-  return f;
-}
-
-bool os::Bsd::libnuma_init() {
-  // sched_getcpu() should be in libc.
-  set_sched_getcpu(CAST_TO_FN_PTR(sched_getcpu_func_t,
-                                  dlsym(RTLD_DEFAULT, "sched_getcpu")));
-
-  if (sched_getcpu() != -1) { // Does it work?
-    void *handle = dlopen("libnuma.so.1", RTLD_LAZY);
-    if (handle != NULL) {
-      set_numa_node_to_cpus(CAST_TO_FN_PTR(numa_node_to_cpus_func_t,
-                                           libnuma_dlsym(handle, "numa_node_to_cpus")));
-      set_numa_max_node(CAST_TO_FN_PTR(numa_max_node_func_t,
-                                       libnuma_dlsym(handle, "numa_max_node")));
-      set_numa_available(CAST_TO_FN_PTR(numa_available_func_t,
-                                        libnuma_dlsym(handle, "numa_available")));
-      set_numa_tonode_memory(CAST_TO_FN_PTR(numa_tonode_memory_func_t,
-                                            libnuma_dlsym(handle, "numa_tonode_memory")));
-      set_numa_interleave_memory(CAST_TO_FN_PTR(numa_interleave_memory_func_t,
-                                            libnuma_dlsym(handle, "numa_interleave_memory")));
-
-
-      if (numa_available() != -1) {
-        set_numa_all_nodes((unsigned long*)libnuma_dlsym(handle, "numa_all_nodes"));
-        // Create a cpu -> node mapping
-        _cpu_to_node = new (ResourceObj::C_HEAP) GrowableArray<int>(0, true);
-        rebuild_cpu_to_node_map();
-        return true;
-      }
-    }
-  }
-  return false;
-}
-
-// rebuild_cpu_to_node_map() constructs a table mapping cpud id to node id.
-// The table is later used in get_node_by_cpu().
-void os::Bsd::rebuild_cpu_to_node_map() {
-  const size_t NCPUS = 32768; // Since the buffer size computation is very obscure
-                              // in libnuma (possible values are starting from 16,
-                              // and continuing up with every other power of 2, but less
-                              // than the maximum number of CPUs supported by kernel), and
-                              // is a subject to change (in libnuma version 2 the requirements
-                              // are more reasonable) we'll just hardcode the number they use
-                              // in the library.
-  const size_t BitsPerCLong = sizeof(long) * CHAR_BIT;
-
-  size_t cpu_num = os::active_processor_count();
-  size_t cpu_map_size = NCPUS / BitsPerCLong;
-  size_t cpu_map_valid_size =
-    MIN2((cpu_num + BitsPerCLong - 1) / BitsPerCLong, cpu_map_size);
-
-  cpu_to_node()->clear();
-  cpu_to_node()->at_grow(cpu_num - 1);
-  size_t node_num = numa_get_groups_num();
-
-  unsigned long *cpu_map = NEW_C_HEAP_ARRAY(unsigned long, cpu_map_size);
-  for (size_t i = 0; i < node_num; i++) {
-    if (numa_node_to_cpus(i, cpu_map, cpu_map_size * sizeof(unsigned long)) != -1) {
-      for (size_t j = 0; j < cpu_map_valid_size; j++) {
-        if (cpu_map[j] != 0) {
-          for (size_t k = 0; k < BitsPerCLong; k++) {
-            if (cpu_map[j] & (1UL << k)) {
-              cpu_to_node()->at_put(j * BitsPerCLong + k, i);
-            }
-          }
-        }
-      }
-    }
-  }
-  FREE_C_HEAP_ARRAY(unsigned long, cpu_map);
-}
-
-int os::Bsd::get_node_by_cpu(int cpu_id) {
-  if (cpu_to_node() != NULL && cpu_id >= 0 && cpu_id < cpu_to_node()->length()) {
-    return cpu_to_node()->at(cpu_id);
-  }
-  return -1;
-}
-
-GrowableArray<int>* os::Bsd::_cpu_to_node;
-os::Bsd::sched_getcpu_func_t os::Bsd::_sched_getcpu;
-os::Bsd::numa_node_to_cpus_func_t os::Bsd::_numa_node_to_cpus;
-os::Bsd::numa_max_node_func_t os::Bsd::_numa_max_node;
-os::Bsd::numa_available_func_t os::Bsd::_numa_available;
-os::Bsd::numa_tonode_memory_func_t os::Bsd::_numa_tonode_memory;
-os::Bsd::numa_interleave_memory_func_t os::Bsd::_numa_interleave_memory;
-unsigned long* os::Bsd::_numa_all_nodes;
-#endif
 
 bool os::pd_uncommit_memory(char* addr, size_t size) {
 #ifdef __OpenBSD__
@@ -3084,42 +2171,7 @@
 }
 
 bool os::Bsd::hugetlbfs_sanity_check(bool warn, size_t page_size) {
-  bool result = false;
-#ifndef _ALLBSD_SOURCE
-  void *p = mmap (NULL, page_size, PROT_READ|PROT_WRITE,
-                  MAP_ANONYMOUS|MAP_PRIVATE|MAP_HUGETLB,
-                  -1, 0);
-
-  if (p != (void *) -1) {
-    // We don't know if this really is a huge page or not.
-    FILE *fp = fopen("/proc/self/maps", "r");
-    if (fp) {
-      while (!feof(fp)) {
-        char chars[257];
-        long x = 0;
-        if (fgets(chars, sizeof(chars), fp)) {
-          if (sscanf(chars, "%lx-%*x", &x) == 1
-              && x == (long)p) {
-            if (strstr (chars, "hugepage")) {
-              result = true;
-              break;
-            }
-          }
-        }
-      }
-      fclose(fp);
-    }
-    munmap (p, page_size);
-    if (result)
-      return true;
-  }
-
-  if (warn) {
-    warning("HugeTLBFS is not supported by the operating system.");
-  }
-#endif
-
-  return result;
+  return false;
 }
 
 /*
@@ -3164,92 +2216,8 @@
 static size_t _large_page_size = 0;
 
 void os::large_page_init() {
-#ifndef _ALLBSD_SOURCE
-  if (!UseLargePages) {
-    UseHugeTLBFS = false;
-    UseSHM = false;
-    return;
-  }
-
-  if (FLAG_IS_DEFAULT(UseHugeTLBFS) && FLAG_IS_DEFAULT(UseSHM)) {
-    // If UseLargePages is specified on the command line try both methods,
-    // if it's default, then try only HugeTLBFS.
-    if (FLAG_IS_DEFAULT(UseLargePages)) {
-      UseHugeTLBFS = true;
-    } else {
-      UseHugeTLBFS = UseSHM = true;
-    }
-  }
-
-  if (LargePageSizeInBytes) {
-    _large_page_size = LargePageSizeInBytes;
-  } else {
-    // large_page_size on Bsd is used to round up heap size. x86 uses either
-    // 2M or 4M page, depending on whether PAE (Physical Address Extensions)
-    // mode is enabled. AMD64/EM64T uses 2M page in 64bit mode. IA64 can use
-    // page as large as 256M.
-    //
-    // Here we try to figure out page size by parsing /proc/meminfo and looking
-    // for a line with the following format:
-    //    Hugepagesize:     2048 kB
-    //
-    // If we can't determine the value (e.g. /proc is not mounted, or the text
-    // format has been changed), we'll use the largest page size supported by
-    // the processor.
-
-#ifndef ZERO
-    _large_page_size = IA32_ONLY(4 * M) AMD64_ONLY(2 * M) IA64_ONLY(256 * M) SPARC_ONLY(4 * M)
-                       ARM_ONLY(2 * M) PPC_ONLY(4 * M);
-#endif // ZERO
-
-    FILE *fp = fopen("/proc/meminfo", "r");
-    if (fp) {
-      while (!feof(fp)) {
-        int x = 0;
-        char buf[16];
-        if (fscanf(fp, "Hugepagesize: %d", &x) == 1) {
-          if (x && fgets(buf, sizeof(buf), fp) && strcmp(buf, " kB\n") == 0) {
-            _large_page_size = x * K;
-            break;
-          }
-        } else {
-          // skip to next line
-          for (;;) {
-            int ch = fgetc(fp);
-            if (ch == EOF || ch == (int)'\n') break;
-          }
-        }
-      }
-      fclose(fp);
-    }
-  }
-
-  // print a warning if any large page related flag is specified on command line
-  bool warn_on_failure = !FLAG_IS_DEFAULT(UseHugeTLBFS);
-
-  const size_t default_page_size = (size_t)Bsd::page_size();
-  if (_large_page_size > default_page_size) {
-    _page_sizes[0] = _large_page_size;
-    _page_sizes[1] = default_page_size;
-    _page_sizes[2] = 0;
-  }
-  UseHugeTLBFS = UseHugeTLBFS &&
-                 Bsd::hugetlbfs_sanity_check(warn_on_failure, _large_page_size);
-
-  if (UseHugeTLBFS)
-    UseSHM = false;
-
-  UseLargePages = UseHugeTLBFS || UseSHM;
-
-  set_coredump_filter();
-#endif
 }
 
-#ifndef _ALLBSD_SOURCE
-#ifndef SHM_HUGETLB
-#define SHM_HUGETLB 04000
-#endif
-#endif
 
 char* os::reserve_memory_special(size_t bytes, char* req_addr, bool exec) {
   // "exec" is passed in but not used.  Creating the shared image for
@@ -3267,11 +2235,7 @@
 
   // Create a large shared memory region to attach to based on size.
   // Currently, size is the total size of the heap
-#ifndef _ALLBSD_SOURCE
-  int shmid = shmget(key, bytes, SHM_HUGETLB|IPC_CREAT|SHM_R|SHM_W);
-#else
   int shmid = shmget(key, bytes, IPC_CREAT|SHM_R|SHM_W);
-#endif
   if (shmid == -1) {
      // Possible reasons for shmget failure:
      // 1. shmmax is too small for Java heap.
@@ -3558,7 +2522,7 @@
 // this reason, the code should not be used as default (ThreadPriorityPolicy=0).
 // It is only used when ThreadPriorityPolicy=1 and requires root privilege.
 
-#if defined(_ALLBSD_SOURCE) && !defined(__APPLE__)
+#if !defined(__APPLE__)
 int os::java_to_os_priority[CriticalPriority + 1] = {
   19,              // 0 Entry should never be used
 
@@ -3578,7 +2542,7 @@
 
   31               // 11 CriticalPriority
 };
-#elif defined(__APPLE__)
+#else
 /* Using Mach high-level priority assignments */
 int os::java_to_os_priority[CriticalPriority + 1] = {
    0,              // 0 Entry should never be used (MINPRI_USER)
@@ -3599,26 +2563,6 @@
 
   36               // 11 CriticalPriority
 };
-#else
-int os::java_to_os_priority[CriticalPriority + 1] = {
-  19,              // 0 Entry should never be used
-
-   4,              // 1 MinPriority
-   3,              // 2
-   2,              // 3
-
-   1,              // 4
-   0,              // 5 NormPriority
-  -1,              // 6
-
-  -2,              // 7
-  -3,              // 8
-  -4,              // 9 NearMaxPriority
-
-  -5,              // 10 MaxPriority
-
-  -5               // 11 CriticalPriority
-};
 #endif
 
 static int prio_init() {
@@ -4179,22 +3123,6 @@
   }
 }
 
-#ifndef _ALLBSD_SOURCE
-// This is the fastest way to get thread cpu time on Bsd.
-// Returns cpu time (user+sys) for any thread, not only for current.
-// POSIX compliant clocks are implemented in the kernels 2.6.16+.
-// It might work on 2.6.10+ with a special kernel/glibc patch.
-// For reference, please, see IEEE Std 1003.1-2004:
-//   http://www.unix.org/single_unix_specification
-
-jlong os::Bsd::fast_thread_cpu_time(clockid_t clockid) {
-  struct timespec tp;
-  int rc = os::Bsd::clock_gettime(clockid, &tp);
-  assert(rc == 0, "clock_gettime is expected to return 0 code");
-
-  return (tp.tv_sec * NANOSECS_PER_SEC) + tp.tv_nsec;
-}
-#endif
 
 /////
 // glibc on Bsd platform uses non-documented flag
@@ -4458,10 +3386,6 @@
 // this is called _after_ the global arguments have been parsed
 jint os::init_2(void)
 {
-#ifndef _ALLBSD_SOURCE
-  Bsd::fast_thread_clock_init();
-#endif
-
   // Allocate a single page and mark it as readable for safepoint polling
   address polling_page = (address) ::mmap(NULL, Bsd::page_size(), PROT_READ, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
   guarantee( polling_page != MAP_FAILED, "os::init_2: failed to allocate polling page" );
@@ -4518,48 +3442,6 @@
   JavaThread::set_stack_size_at_create(round_to(threadStackSizeInBytes,
         vm_page_size()));
 
-#ifndef _ALLBSD_SOURCE
-  Bsd::capture_initial_stack(JavaThread::stack_size_at_create());
-
-  Bsd::libpthread_init();
-  if (PrintMiscellaneous && (Verbose || WizardMode)) {
-     tty->print_cr("[HotSpot is running with %s, %s(%s)]\n",
-          Bsd::glibc_version(), Bsd::libpthread_version(),
-          Bsd::is_floating_stack() ? "floating stack" : "fixed stack");
-  }
-
-  if (UseNUMA) {
-    if (!Bsd::libnuma_init()) {
-      UseNUMA = false;
-    } else {
-      if ((Bsd::numa_max_node() < 1)) {
-        // There's only one node(they start from 0), disable NUMA.
-        UseNUMA = false;
-      }
-    }
-    // With SHM large pages we cannot uncommit a page, so there's not way
-    // we can make the adaptive lgrp chunk resizing work. If the user specified
-    // both UseNUMA and UseLargePages (or UseSHM) on the command line - warn and
-    // disable adaptive resizing.
-    if (UseNUMA && UseLargePages && UseSHM) {
-      if (!FLAG_IS_DEFAULT(UseNUMA)) {
-        if (FLAG_IS_DEFAULT(UseLargePages) && FLAG_IS_DEFAULT(UseSHM)) {
-          UseLargePages = false;
-        } else {
-          warning("UseNUMA is not fully compatible with SHM large pages, disabling adaptive resizing");
-          UseAdaptiveSizePolicy = false;
-          UseAdaptiveNUMAChunkSizing = false;
-        }
-      } else {
-        UseNUMA = false;
-      }
-    }
-    if (!UseNUMA && ForceNUMA) {
-      UseNUMA = true;
-    }
-  }
-#endif
-
   if (MaxFDLimit) {
     // set the number of file descriptors to max. print out error
     // if getrlimit/setrlimit fails but continue regardless.
@@ -4586,11 +3468,6 @@
     }
   }
 
-#ifndef _ALLBSD_SOURCE
-  // Initialize lock used to serialize thread creation (see os::create_thread)
-  Bsd::set_createThread_lock(new Mutex(Mutex::leaf, "createThread_lock", false));
-#endif
-
   // at-exit methods are called in the reverse order of their registration.
   // atexit functions are called on return from main or as a result of a
   // call to exit(3C). There can be only 32 of these functions registered
@@ -4641,15 +3518,7 @@
 };
 
 int os::active_processor_count() {
-#ifdef _ALLBSD_SOURCE
   return _processor_count;
-#else
-  // Bsd doesn't yet have a (official) notion of processor sets,
-  // so just return the number of online processors.
-  int online_cpus = ::sysconf(_SC_NPROCESSORS_ONLN);
-  assert(online_cpus > 0 && online_cpus <= processor_count(), "sanity check");
-  return online_cpus;
-#endif
 }
 
 void os::set_native_thread_name(const char *name) {
@@ -4703,25 +3572,7 @@
 
 int os::Bsd::safe_cond_timedwait(pthread_cond_t *_cond, pthread_mutex_t *_mutex, const struct timespec *_abstime)
 {
-#ifdef _ALLBSD_SOURCE
   return pthread_cond_timedwait(_cond, _mutex, _abstime);
-#else
-   if (is_NPTL()) {
-      return pthread_cond_timedwait(_cond, _mutex, _abstime);
-   } else {
-#ifndef IA64
-      // 6292965: BsdThreads pthread_cond_timedwait() resets FPU control
-      // word back to default 64bit precision if condvar is signaled. Java
-      // wants 53bit precision.  Save and restore current value.
-      int fpu = get_fpu_control_word();
-#endif // IA64
-      int status = pthread_cond_timedwait(_cond, _mutex, _abstime);
-#ifndef IA64
-      set_fpu_control_word(fpu);
-#endif // IA64
-      return status;
-   }
-#endif
 }
 
 ////////////////////////////////////////////////////////////////////////////////
@@ -5041,20 +3892,6 @@
   return munmap(addr, bytes) == 0;
 }
 
-#ifndef _ALLBSD_SOURCE
-static jlong slow_thread_cpu_time(Thread *thread, bool user_sys_cpu_time);
-
-static clockid_t thread_cpu_clockid(Thread* thread) {
-  pthread_t tid = thread->osthread()->pthread_id();
-  clockid_t clockid;
-
-  // Get thread clockid
-  int rc = os::Bsd::pthread_getcpuclockid(tid, &clockid);
-  assert(rc == 0, "pthread_getcpuclockid is expected to return 0 code");
-  return clockid;
-}
-#endif
-
 // current_thread_cpu_time(bool) and thread_cpu_time(Thread*, bool)
 // are used by JVM M&M and JVMTI to get user+sys or user CPU time
 // of a thread.
@@ -5065,36 +3902,15 @@
 jlong os::current_thread_cpu_time() {
 #ifdef __APPLE__
   return os::thread_cpu_time(Thread::current(), true /* user + sys */);
-#elif !defined(_ALLBSD_SOURCE)
-  if (os::Bsd::supports_fast_thread_cpu_time()) {
-    return os::Bsd::fast_thread_cpu_time(CLOCK_THREAD_CPUTIME_ID);
-  } else {
-    // return user + sys since the cost is the same
-    return slow_thread_cpu_time(Thread::current(), true /* user + sys */);
-  }
 #endif
 }
 
 jlong os::thread_cpu_time(Thread* thread) {
-#ifndef _ALLBSD_SOURCE
-  // consistent with what current_thread_cpu_time() returns
-  if (os::Bsd::supports_fast_thread_cpu_time()) {
-    return os::Bsd::fast_thread_cpu_time(thread_cpu_clockid(thread));
-  } else {
-    return slow_thread_cpu_time(thread, true /* user + sys */);
-  }
-#endif
 }
 
 jlong os::current_thread_cpu_time(bool user_sys_cpu_time) {
 #ifdef __APPLE__
   return os::thread_cpu_time(Thread::current(), user_sys_cpu_time);
-#elif !defined(_ALLBSD_SOURCE)
-  if (user_sys_cpu_time && os::Bsd::supports_fast_thread_cpu_time()) {
-    return os::Bsd::fast_thread_cpu_time(CLOCK_THREAD_CPUTIME_ID);
-  } else {
-    return slow_thread_cpu_time(Thread::current(), user_sys_cpu_time);
-  }
 #endif
 }
 
@@ -5118,106 +3934,9 @@
   } else {
     return ((jlong)tinfo.user_time.seconds * 1000000000) + ((jlong)tinfo.user_time.microseconds * (jlong)1000);
   }
-#elif !defined(_ALLBSD_SOURCE)
-  if (user_sys_cpu_time && os::Bsd::supports_fast_thread_cpu_time()) {
-    return os::Bsd::fast_thread_cpu_time(thread_cpu_clockid(thread));
-  } else {
-    return slow_thread_cpu_time(thread, user_sys_cpu_time);
-  }
 #endif
 }
 
-#ifndef _ALLBSD_SOURCE
-//
-//  -1 on error.
-//
-
-static jlong slow_thread_cpu_time(Thread *thread, bool user_sys_cpu_time) {
-  static bool proc_pid_cpu_avail = true;
-  static bool proc_task_unchecked = true;
-  static const char *proc_stat_path = "/proc/%d/stat";
-  pid_t  tid = thread->osthread()->thread_id();
-  int i;
-  char *s;
-  char stat[2048];
-  int statlen;
-  char proc_name[64];
-  int count;
-  long sys_time, user_time;
-  char string[64];
-  char cdummy;
-  int idummy;
-  long ldummy;
-  FILE *fp;
-
-  // We first try accessing /proc/<pid>/cpu since this is faster to
-  // process.  If this file is not present (bsd kernels 2.5 and above)
-  // then we open /proc/<pid>/stat.
-  if ( proc_pid_cpu_avail ) {
-    sprintf(proc_name, "/proc/%d/cpu", tid);
-    fp =  fopen(proc_name, "r");
-    if ( fp != NULL ) {
-      count = fscanf( fp, "%s %lu %lu\n", string, &user_time, &sys_time);
-      fclose(fp);
-      if ( count != 3 ) return -1;
-
-      if (user_sys_cpu_time) {
-        return ((jlong)sys_time + (jlong)user_time) * (1000000000 / clock_tics_per_sec);
-      } else {
-        return (jlong)user_time * (1000000000 / clock_tics_per_sec);
-      }
-    }
-    else proc_pid_cpu_avail = false;
-  }
-
-  // The /proc/<tid>/stat aggregates per-process usage on
-  // new Bsd kernels 2.6+ where NPTL is supported.
-  // The /proc/self/task/<tid>/stat still has the per-thread usage.
-  // See bug 6328462.
-  // There can be no directory /proc/self/task on kernels 2.4 with NPTL
-  // and possibly in some other cases, so we check its availability.
-  if (proc_task_unchecked && os::Bsd::is_NPTL()) {
-    // This is executed only once
-    proc_task_unchecked = false;
-    fp = fopen("/proc/self/task", "r");
-    if (fp != NULL) {
-      proc_stat_path = "/proc/self/task/%d/stat";
-      fclose(fp);
-    }
-  }
-
-  sprintf(proc_name, proc_stat_path, tid);
-  fp = fopen(proc_name, "r");
-  if ( fp == NULL ) return -1;
-  statlen = fread(stat, 1, 2047, fp);
-  stat[statlen] = '\0';
-  fclose(fp);
-
-  // Skip pid and the command string. Note that we could be dealing with
-  // weird command names, e.g. user could decide to rename java launcher
-  // to "java 1.4.2 :)", then the stat file would look like
-  //                1234 (java 1.4.2 :)) R ... ...
-  // We don't really need to know the command string, just find the last
-  // occurrence of ")" and then start parsing from there. See bug 4726580.
-  s = strrchr(stat, ')');
-  i = 0;
-  if (s == NULL ) return -1;
-
-  // Skip blank chars
-  do s++; while (isspace(*s));
-
-  count = sscanf(s,"%c %d %d %d %d %d %lu %lu %lu %lu %lu %lu %lu",
-                 &cdummy, &idummy, &idummy, &idummy, &idummy, &idummy,
-                 &ldummy, &ldummy, &ldummy, &ldummy, &ldummy,
-                 &user_time, &sys_time);
-  if ( count != 13 ) return -1;
-  if (user_sys_cpu_time) {
-    return ((jlong)sys_time + (jlong)user_time) * (1000000000 / clock_tics_per_sec);
-  } else {
-    return (jlong)user_time * (1000000000 / clock_tics_per_sec);
-  }
-}
-#endif
 
 void os::current_thread_cpu_time_info(jvmtiTimerInfo *info_ptr) {
   info_ptr->max_value = ALL_64_BITS;       // will not wrap in less than 64 bits
@@ -5236,10 +3955,8 @@
 bool os::is_thread_cpu_time_supported() {
 #ifdef __APPLE__
   return true;
-#elif defined(_ALLBSD_SOURCE)
+#else
   return false;
-#else
-  return true;
 #endif
 }