Vendor import of llvm 3.0 final release:

http://llvm.org/svn/llvm-project/llvm/tags/RELEASE_30/final@145349

22 files changed, 407 insertions, 161 deletions

diff --git a/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp b/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
index 88b7524..6fe476d 100644
--- a/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
+++ b/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp
@@ -527,18 +527,20 @@ bool CompileUnit::addConstantValue(DIE *Die, const ConstantInt *CI,
 
   // Get the raw data form of the large APInt.
   const APInt Val = CI->getValue();
-  const char *Ptr = (const char*)Val.getRawData();
+  const uint64_t *Ptr64 = Val.getRawData();
 
   int NumBytes = Val.getBitWidth() / 8; // 8 bits per byte.
   bool LittleEndian = Asm->getTargetData().isLittleEndian();
-  int Incr = (LittleEndian ? 1 : -1);
-  int Start = (LittleEndian ? 0 : NumBytes - 1);
-  int Stop = (LittleEndian ? NumBytes : -1);
 
   // Output the constant to DWARF one byte at a time.
-  for (; Start != Stop; Start += Incr)
-    addUInt(Block, 0, dwarf::DW_FORM_data1,
-            (unsigned char)0xFF & Ptr[Start]);
+  for (int i = 0; i < NumBytes; i++) {
+    uint8_t c;
+    if (LittleEndian)
+      c = Ptr64[i / 8] >> (8 * (i & 7));
+    else
+      c = Ptr64[(NumBytes - 1 - i) / 8] >> (8 * ((NumBytes - 1 - i) & 7));
+    addUInt(Block, 0, dwarf::DW_FORM_data1, c);
+  }
 
   addBlock(Die, dwarf::DW_AT_const_value, 0, Block);
   return true;
diff --git a/lib/CodeGen/LLVMTargetMachine.cpp b/lib/CodeGen/LLVMTargetMachine.cpp
index 80ecc22..187147a 100644
--- a/lib/CodeGen/LLVMTargetMachine.cpp
+++ b/lib/CodeGen/LLVMTargetMachine.cpp
@@ -119,7 +119,8 @@ LLVMTargetMachine::LLVMTargetMachine(const Target &T, StringRef Triple,
   // we'll crash later.
   // Provide the user with a useful error message about what's wrong.
   assert(AsmInfo && "MCAsmInfo not initialized."
-	 "Make sure you include the correct TargetSelect.h!");
+         "Make sure you include the correct TargetSelect.h"
+         "and that InitializeAllTargetMCs() is being invoked!");
 }
 
 bool LLVMTargetMachine::addPassesToEmitFile(PassManagerBase &PM,
diff --git a/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp b/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
index 7ed46a6..095b400 100644
--- a/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
+++ b/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
@@ -2034,14 +2034,17 @@ bool SelectionDAGBuilder::handleJTSwitchCase(CaseRec &CR,
     return false;
 
   APInt Range = ComputeRange(First, Last);
-  double Density = TSize.roundToDouble() / Range.roundToDouble();
-  if (Density < 0.4)
+  // The density is TSize / Range. Require at least 40%.
+  // It should not be possible for IntTSize to saturate for sane code, but make
+  // sure we handle Range saturation correctly.
+  uint64_t IntRange = Range.getLimitedValue(UINT64_MAX/10);
+  uint64_t IntTSize = TSize.getLimitedValue(UINT64_MAX/10);
+  if (IntTSize * 10 < IntRange * 4)
     return false;
 
   DEBUG(dbgs() << "Lowering jump table\n"
                << "First entry: " << First << ". Last entry: " << Last << '\n'
-               << "Range: " << Range
-               << ". Size: " << TSize << ". Density: " << Density << "\n\n");
+               << "Range: " << Range << ". Size: " << TSize << ".\n\n");
 
   // Get the MachineFunction which holds the current MBB.  This is used when
   // inserting any additional MBBs necessary to represent the switch.
diff --git a/lib/CodeGen/TargetLoweringObjectFileImpl.cpp b/lib/CodeGen/TargetLoweringObjectFileImpl.cpp
index fb87154..3848f4d 100644
--- a/lib/CodeGen/TargetLoweringObjectFileImpl.cpp
+++ b/lib/CodeGen/TargetLoweringObjectFileImpl.cpp
@@ -506,7 +506,9 @@ getExprForDwarfGlobalReference(const GlobalValue *GV, Mangler *Mang,
     // Add information about the stub reference to MachOMMI so that the stub
     // gets emitted by the asmprinter.
     MCSymbol *SSym = getContext().GetOrCreateSymbol(Name.str());
-    MachineModuleInfoImpl::StubValueTy &StubSym = MachOMMI.getGVStubEntry(SSym);
+    MachineModuleInfoImpl::StubValueTy &StubSym =
+      GV->hasHiddenVisibility() ? MachOMMI.getHiddenGVStubEntry(SSym) :
+                                  MachOMMI.getGVStubEntry(SSym);
     if (StubSym.getPointer() == 0) {
       MCSymbol *Sym = Mang->getSymbol(GV);
       StubSym = MachineModuleInfoImpl::StubValueTy(Sym, !GV->hasLocalLinkage());
@@ -534,7 +536,9 @@ getCFIPersonalitySymbol(const GlobalValue *GV, Mangler *Mang,
   // Add information about the stub reference to MachOMMI so that the stub
   // gets emitted by the asmprinter.
   MCSymbol *SSym = getContext().GetOrCreateSymbol(Name.str());
-  MachineModuleInfoImpl::StubValueTy &StubSym = MachOMMI.getGVStubEntry(SSym);
+  MachineModuleInfoImpl::StubValueTy &StubSym =
+      GV->hasHiddenVisibility() ? MachOMMI.getHiddenGVStubEntry(SSym) :
+                                  MachOMMI.getGVStubEntry(SSym);
   if (StubSym.getPointer() == 0) {
     MCSymbol *Sym = Mang->getSymbol(GV);
     StubSym = MachineModuleInfoImpl::StubValueTy(Sym, !GV->hasLocalLinkage());
diff --git a/lib/Target/ARM/ARMBaseRegisterInfo.cpp b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
index 7c42342..154f1f8 100644
--- a/lib/Target/ARM/ARMBaseRegisterInfo.cpp
+++ b/lib/Target/ARM/ARMBaseRegisterInfo.cpp
@@ -63,6 +63,13 @@ ARMBaseRegisterInfo::ARMBaseRegisterInfo(const ARMBaseInstrInfo &tii,
 
 const unsigned*
 ARMBaseRegisterInfo::getCalleeSavedRegs(const MachineFunction *MF) const {
+  bool ghcCall = false;
+
+  if (MF) {
+    const Function *F = MF->getFunction();
+    ghcCall = (F ? F->getCallingConv() == CallingConv::GHC : false);
+  }
+
   static const unsigned CalleeSavedRegs[] = {
     ARM::LR, ARM::R11, ARM::R10, ARM::R9, ARM::R8,
     ARM::R7, ARM::R6,  ARM::R5,  ARM::R4,
@@ -82,7 +89,13 @@ ARMBaseRegisterInfo::getCalleeSavedRegs(const MachineFunction *MF) const {
     ARM::D11, ARM::D10, ARM::D9,  ARM::D8,
     0
   };
-  return STI.isTargetDarwin() ? DarwinCalleeSavedRegs : CalleeSavedRegs;
+
+  static const unsigned GhcCalleeSavedRegs[] = {
+    0
+  };
+
+  return ghcCall ? GhcCalleeSavedRegs :
+         STI.isTargetDarwin() ? DarwinCalleeSavedRegs : CalleeSavedRegs;
 }
 
 BitVector ARMBaseRegisterInfo::
diff --git a/lib/Target/ARM/ARMCallingConv.td b/lib/Target/ARM/ARMCallingConv.td
index d2981c0..47b2e98 100644
--- a/lib/Target/ARM/ARMCallingConv.td
+++ b/lib/Target/ARM/ARMCallingConv.td
@@ -82,6 +82,25 @@ def RetFastCC_ARM_APCS : CallingConv<[
   CCDelegateTo<RetCC_ARM_APCS>
 ]>;
 
+//===----------------------------------------------------------------------===//
+// ARM APCS Calling Convention for GHC
+//===----------------------------------------------------------------------===//
+
+def CC_ARM_APCS_GHC : CallingConv<[
+  // Handle all vector types as either f64 or v2f64.
+  CCIfType<[v1i64, v2i32, v4i16, v8i8, v2f32], CCBitConvertToType<f64>>,
+  CCIfType<[v2i64, v4i32, v8i16, v16i8, v4f32], CCBitConvertToType<v2f64>>,
+
+  CCIfType<[v2f64], CCAssignToReg<[Q4, Q5]>>,
+  CCIfType<[f64], CCAssignToReg<[D8, D9, D10, D11]>>,
+  CCIfType<[f32], CCAssignToReg<[S16, S17, S18, S19, S20, S21, S22, S23]>>,
+
+  // Promote i8/i16 arguments to i32.
+  CCIfType<[i8, i16], CCPromoteToType<i32>>,
+
+  // Pass in STG registers: Base, Sp, Hp, R1, R2, R3, R4, SpLim
+  CCIfType<[i32], CCAssignToReg<[R4, R5, R6, R7, R8, R9, R10, R11]>>
+]>;
 
 //===----------------------------------------------------------------------===//
 // ARM AAPCS (EABI) Calling Convention, common parts
diff --git a/lib/Target/ARM/ARMFastISel.cpp b/lib/Target/ARM/ARMFastISel.cpp
index 9bc7ef2..dc8e54d 100644
--- a/lib/Target/ARM/ARMFastISel.cpp
+++ b/lib/Target/ARM/ARMFastISel.cpp
@@ -1548,6 +1548,11 @@ CCAssignFn *ARMFastISel::CCAssignFnForCall(CallingConv::ID CC, bool Return) {
     return (Return ? RetCC_ARM_AAPCS: CC_ARM_AAPCS);
   case CallingConv::ARM_APCS:
     return (Return ? RetCC_ARM_APCS: CC_ARM_APCS);
+  case CallingConv::GHC:
+    if (Return)
+      llvm_unreachable("Can't return in GHC call convention");
+    else
+      return CC_ARM_APCS_GHC;
   }
 }
 
diff --git a/lib/Target/ARM/ARMFrameLowering.cpp b/lib/Target/ARM/ARMFrameLowering.cpp
index 2d1de6f..412751b 100644
--- a/lib/Target/ARM/ARMFrameLowering.cpp
+++ b/lib/Target/ARM/ARMFrameLowering.cpp
@@ -15,6 +15,8 @@
 #include "ARMBaseInstrInfo.h"
 #include "ARMBaseRegisterInfo.h"
 #include "ARMMachineFunctionInfo.h"
+#include "llvm/CallingConv.h"
+#include "llvm/Function.h"
 #include "MCTargetDesc/ARMAddressingModes.h"
 #include "llvm/CodeGen/MachineFrameInfo.h"
 #include "llvm/CodeGen/MachineFunction.h"
@@ -139,6 +141,10 @@ void ARMFrameLowering::emitPrologue(MachineFunction &MF) const {
   unsigned GPRCS1Size = 0, GPRCS2Size = 0, DPRCSSize = 0;
   int FramePtrSpillFI = 0;
 
+  // All calls are tail calls in GHC calling conv, and functions have no prologue/epilogue.
+  if (MF.getFunction()->getCallingConv() == CallingConv::GHC)
+    return;
+
   // Allocate the vararg register save area. This is not counted in NumBytes.
   if (VARegSaveSize)
     emitSPUpdate(isARM, MBB, MBBI, dl, TII, -VARegSaveSize,
@@ -326,6 +332,10 @@ void ARMFrameLowering::emitEpilogue(MachineFunction &MF,
   int NumBytes = (int)MFI->getStackSize();
   unsigned FramePtr = RegInfo->getFrameRegister(MF);
 
+  // All calls are tail calls in GHC calling conv, and functions have no prologue/epilogue.
+  if (MF.getFunction()->getCallingConv() == CallingConv::GHC)
+    return;
+
   if (!AFI->hasStackFrame()) {
     if (NumBytes != 0)
       emitSPUpdate(isARM, MBB, MBBI, dl, TII, NumBytes);
diff --git a/lib/Target/ARM/ARMISelLowering.cpp b/lib/Target/ARM/ARMISelLowering.cpp
index e44e356..f60d177 100644
--- a/lib/Target/ARM/ARMISelLowering.cpp
+++ b/lib/Target/ARM/ARMISelLowering.cpp
@@ -1091,6 +1091,8 @@ CCAssignFn *ARMTargetLowering::CCAssignFnForNode(CallingConv::ID CC,
     return (Return ? RetCC_ARM_AAPCS : CC_ARM_AAPCS);
   case CallingConv::ARM_APCS:
     return (Return ? RetCC_ARM_APCS : CC_ARM_APCS);
+  case CallingConv::GHC:
+    return (Return ? RetCC_ARM_APCS : CC_ARM_APCS_GHC);
   }
 }
 
diff --git a/lib/Target/ARM/ARMInstrThumb2.td b/lib/Target/ARM/ARMInstrThumb2.td
index 471ec29..05dcc89 100644
--- a/lib/Target/ARM/ARMInstrThumb2.td
+++ b/lib/Target/ARM/ARMInstrThumb2.td
@@ -1538,8 +1538,7 @@ multiclass thumb2_ld_mult<string asm, InstrItinClass itin,
     let Inst{21}    = 0;        // No writeback
     let Inst{20}    = L_bit;
     let Inst{19-16} = Rn;
-    let Inst{15}    = 0;
-    let Inst{14-0}  = regs{14-0};
+    let Inst{15-0}  = regs;
   }
   def IA_UPD :
     T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops),
@@ -1554,8 +1553,7 @@ multiclass thumb2_ld_mult<string asm, InstrItinClass itin,
     let Inst{21}    = 1;        // Writeback
     let Inst{20}    = L_bit;
     let Inst{19-16} = Rn;
-    let Inst{15}    = 0;
-    let Inst{14-0}  = regs{14-0};
+    let Inst{15-0}  = regs;
   }
   def DB :
     T2XI<(outs), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops),
@@ -1570,8 +1568,7 @@ multiclass thumb2_ld_mult<string asm, InstrItinClass itin,
     let Inst{21}    = 0;        // No writeback
     let Inst{20}    = L_bit;
     let Inst{19-16} = Rn;
-    let Inst{15}    = 0;
-    let Inst{14-0}  = regs{14-0};
+    let Inst{15-0}  = regs;
   }
   def DB_UPD :
     T2XIt<(outs GPR:$wb), (ins GPR:$Rn, pred:$p, reglist:$regs, variable_ops),
@@ -1586,8 +1583,7 @@ multiclass thumb2_ld_mult<string asm, InstrItinClass itin,
     let Inst{21}    = 1;        // Writeback
     let Inst{20}    = L_bit;
     let Inst{19-16} = Rn;
-    let Inst{15}    = 0;
-    let Inst{14-0}  = regs{14-0};
+    let Inst{15-0}  = regs;
   }
 }
 
diff --git a/lib/Target/CppBackend/CPPBackend.cpp b/lib/Target/CppBackend/CPPBackend.cpp
index ae0e3c4..394ea2b 100644
--- a/lib/Target/CppBackend/CPPBackend.cpp
+++ b/lib/Target/CppBackend/CPPBackend.cpp
@@ -1016,6 +1016,27 @@ std::string CppWriter::getOpName(const Value* V) {
   return result;
 }
 
+static StringRef ConvertAtomicOrdering(AtomicOrdering Ordering) {
+  switch (Ordering) {
+    case NotAtomic: return "NotAtomic";
+    case Unordered: return "Unordered";
+    case Monotonic: return "Monotonic";
+    case Acquire: return "Acquire";
+    case Release: return "Release";
+    case AcquireRelease: return "AcquireRelease";
+    case SequentiallyConsistent: return "SequentiallyConsistent";
+  }
+  llvm_unreachable("Unknown ordering");
+}
+
+static StringRef ConvertAtomicSynchScope(SynchronizationScope SynchScope) {
+  switch (SynchScope) {
+    case SingleThread: return "SingleThread";
+    case CrossThread: return "CrossThread";
+  }
+  llvm_unreachable("Unknown synch scope");
+}
+
 // printInstruction - This member is called for each Instruction in a function.
 void CppWriter::printInstruction(const Instruction *I,
                                  const std::string& bbname) {
@@ -1237,15 +1258,33 @@ void CppWriter::printInstruction(const Instruction *I,
     printEscapedString(load->getName());
     Out << "\", " << (load->isVolatile() ? "true" : "false" )
         << ", " << bbname << ");";
+    if (load->getAlignment())
+      nl(Out) << iName << "->setAlignment("
+              << load->getAlignment() << ");";
+    if (load->isAtomic()) {
+      StringRef Ordering = ConvertAtomicOrdering(load->getOrdering());
+      StringRef CrossThread = ConvertAtomicSynchScope(load->getSynchScope());
+      nl(Out) << iName << "->setAtomic("
+              << Ordering << ", " << CrossThread << ");";
+    }
     break;
   }
   case Instruction::Store: {
     const StoreInst* store = cast<StoreInst>(I);
-    Out << " new StoreInst("
+    Out << "StoreInst* " << iName << " = new StoreInst("
         << opNames[0] << ", "
         << opNames[1] << ", "
         << (store->isVolatile() ? "true" : "false")
         << ", " << bbname << ");";
+    if (store->getAlignment())
+      nl(Out) << iName << "->setAlignment("
+              << store->getAlignment() << ");";
+    if (store->isAtomic()) {
+      StringRef Ordering = ConvertAtomicOrdering(store->getOrdering());
+      StringRef CrossThread = ConvertAtomicSynchScope(store->getSynchScope());
+      nl(Out) << iName << "->setAtomic("
+              << Ordering << ", " << CrossThread << ");";
+    }
     break;
   }
   case Instruction::GetElementPtr: {
@@ -1447,6 +1486,60 @@ void CppWriter::printInstruction(const Instruction *I,
     Out << "\", " << bbname << ");";
     break;
   }
+  case Instruction::Fence: {
+    const FenceInst *fi = cast<FenceInst>(I);
+    StringRef Ordering = ConvertAtomicOrdering(fi->getOrdering());
+    StringRef CrossThread = ConvertAtomicSynchScope(fi->getSynchScope());
+    Out << "FenceInst* " << iName
+        << " = new FenceInst(mod->getContext(), "
+        << Ordering << ", " << CrossThread << ", " << bbname
+        << ");";
+    break;
+  }
+  case Instruction::AtomicCmpXchg: {
+    const AtomicCmpXchgInst *cxi = cast<AtomicCmpXchgInst>(I);
+    StringRef Ordering = ConvertAtomicOrdering(cxi->getOrdering());
+    StringRef CrossThread = ConvertAtomicSynchScope(cxi->getSynchScope());
+    Out << "AtomicCmpXchgInst* " << iName
+        << " = new AtomicCmpXchgInst("
+        << opNames[0] << ", " << opNames[1] << ", " << opNames[2] << ", "
+        << Ordering << ", " << CrossThread << ", " << bbname
+        << ");";
+    nl(Out) << iName << "->setName(\"";
+    printEscapedString(cxi->getName());
+    Out << "\");";
+    break;
+  }
+  case Instruction::AtomicRMW: {
+    const AtomicRMWInst *rmwi = cast<AtomicRMWInst>(I);
+    StringRef Ordering = ConvertAtomicOrdering(rmwi->getOrdering());
+    StringRef CrossThread = ConvertAtomicSynchScope(rmwi->getSynchScope());
+    StringRef Operation;
+    switch (rmwi->getOperation()) {
+      case AtomicRMWInst::Xchg: Operation = "AtomicRMWInst::Xchg"; break;
+      case AtomicRMWInst::Add:  Operation = "AtomicRMWInst::Add"; break;
+      case AtomicRMWInst::Sub:  Operation = "AtomicRMWInst::Sub"; break;
+      case AtomicRMWInst::And:  Operation = "AtomicRMWInst::And"; break;
+      case AtomicRMWInst::Nand: Operation = "AtomicRMWInst::Nand"; break;
+      case AtomicRMWInst::Or:   Operation = "AtomicRMWInst::Or"; break;
+      case AtomicRMWInst::Xor:  Operation = "AtomicRMWInst::Xor"; break;
+      case AtomicRMWInst::Max:  Operation = "AtomicRMWInst::Max"; break;
+      case AtomicRMWInst::Min:  Operation = "AtomicRMWInst::Min"; break;
+      case AtomicRMWInst::UMax: Operation = "AtomicRMWInst::UMax"; break;
+      case AtomicRMWInst::UMin: Operation = "AtomicRMWInst::UMin"; break;
+      case AtomicRMWInst::BAD_BINOP: llvm_unreachable("Bad atomic operation");
+    }
+    Out << "AtomicRMWInst* " << iName
+        << " = new AtomicRMWInst("
+        << Operation << ", "
+        << opNames[0] << ", " << opNames[1] << ", "
+        << Ordering << ", " << CrossThread << ", " << bbname
+        << ");";
+    nl(Out) << iName << "->setName(\"";
+    printEscapedString(rmwi->getName());
+    Out << "\");";
+    break;
+  }
   }
   DefinedValues.insert(I);
   nl(Out);
@@ -1623,7 +1716,9 @@ void CppWriter::printFunctionBody(const Function *F) {
       Out << "Value* " << getCppName(AI) << " = args++;";
       nl(Out);
       if (AI->hasName()) {
-        Out << getCppName(AI) << "->setName(\"" << AI->getName() << "\");";
+        Out << getCppName(AI) << "->setName(\"";
+        printEscapedString(AI->getName());
+        Out << "\");";
         nl(Out);
       }
     }
diff --git a/lib/Target/Mips/CMakeLists.txt b/lib/Target/Mips/CMakeLists.txt
index 1b4329b..71391f3 100644
--- a/lib/Target/Mips/CMakeLists.txt
+++ b/lib/Target/Mips/CMakeLists.txt
@@ -2,6 +2,7 @@ set(LLVM_TARGET_DEFINITIONS Mips.td)
 
 llvm_tablegen(MipsGenRegisterInfo.inc -gen-register-info)
 llvm_tablegen(MipsGenInstrInfo.inc -gen-instr-info)
+llvm_tablegen(MipsGenCodeEmitter.inc -gen-emitter)
 llvm_tablegen(MipsGenAsmWriter.inc -gen-asm-writer)
 llvm_tablegen(MipsGenDAGISel.inc -gen-dag-isel)
 llvm_tablegen(MipsGenCallingConv.inc -gen-callingconv)
diff --git a/lib/Target/Mips/Makefile b/lib/Target/Mips/Makefile
index cc4a8ae..d72693c 100644
--- a/lib/Target/Mips/Makefile
+++ b/lib/Target/Mips/Makefile
@@ -13,7 +13,7 @@ TARGET = Mips
 
 # Make sure that tblgen is run, first thing.
 BUILT_SOURCES = MipsGenRegisterInfo.inc MipsGenInstrInfo.inc \
-                MipsGenAsmWriter.inc \
+                MipsGenAsmWriter.inc MipsGenCodeEmitter.inc \
                 MipsGenDAGISel.inc MipsGenCallingConv.inc \
                 MipsGenSubtargetInfo.inc
 
diff --git a/lib/Target/Mips/Mips64InstrInfo.td b/lib/Target/Mips/Mips64InstrInfo.td
index 49b0223..9758f4b 100644
--- a/lib/Target/Mips/Mips64InstrInfo.td
+++ b/lib/Target/Mips/Mips64InstrInfo.td
@@ -39,51 +39,51 @@ def imm32_63 : ImmLeaf<i64,
 // Shifts
 class LogicR_shift_rotate_imm64<bits<6> func, bits<5> _rs, string instr_asm,
                                 SDNode OpNode, PatFrag PF>:
-  FR<0x00, func, (outs CPU64Regs:$dst), (ins CPU64Regs:$b, shamt_64:$c),
-     !strconcat(instr_asm, "\t$dst, $b, $c"),
-     [(set CPU64Regs:$dst, (OpNode CPU64Regs:$b, (i64 PF:$c)))],
+  FR<0x00, func, (outs CPU64Regs:$rd), (ins CPU64Regs:$rt, shamt_64:$shamt),
+     !strconcat(instr_asm, "\t$rd, $rt, $shamt"),
+     [(set CPU64Regs:$rd, (OpNode CPU64Regs:$rt, (i64 PF:$shamt)))],
      IIAlu> {
   let rs = _rs;
 }
 
 class LogicR_shift_rotate_reg64<bits<6> func, bits<5> _shamt, string instr_asm,
                                 SDNode OpNode>:
-  FR<0x00, func, (outs CPU64Regs:$dst), (ins CPU64Regs:$c, CPU64Regs:$b),
-     !strconcat(instr_asm, "\t$dst, $b, $c"),
-     [(set CPU64Regs:$dst, (OpNode CPU64Regs:$b, CPU64Regs:$c))], IIAlu> {
+  FR<0x00, func, (outs CPU64Regs:$rd), (ins CPU64Regs:$rs, CPU64Regs:$rt),
+     !strconcat(instr_asm, "\t$rd, $rt, $rs"),
+     [(set CPU64Regs:$rd, (OpNode CPU64Regs:$rt, CPU64Regs:$rs))], IIAlu> {
   let shamt = _shamt;
 }
 
 // Mul, Div
-let Defs = [HI64, LO64] in {
+let rd = 0, shamt = 0, Defs = [HI64, LO64] in {
   let isCommutable = 1 in
   class Mul64<bits<6> func, string instr_asm, InstrItinClass itin>:
-    FR<0x00, func, (outs), (ins CPU64Regs:$a, CPU64Regs:$b),
-       !strconcat(instr_asm, "\t$a, $b"), [], itin>;
+    FR<0x00, func, (outs), (ins CPU64Regs:$rs, CPU64Regs:$rt),
+       !strconcat(instr_asm, "\t$rs, $rt"), [], itin>;
 
   class Div64<SDNode op, bits<6> func, string instr_asm, InstrItinClass itin>:
-              FR<0x00, func, (outs), (ins CPU64Regs:$a, CPU64Regs:$b),
-              !strconcat(instr_asm, "\t$$zero, $a, $b"),
-              [(op CPU64Regs:$a, CPU64Regs:$b)], itin>;
+              FR<0x00, func, (outs), (ins CPU64Regs:$rs, CPU64Regs:$rt),
+              !strconcat(instr_asm, "\t$$zero, $rs, $rt"),
+              [(op CPU64Regs:$rs, CPU64Regs:$rt)], itin>;
 }
 
 // Move from Hi/Lo
 let shamt = 0 in {
 let rs = 0, rt = 0 in
 class MoveFromLOHI64<bits<6> func, string instr_asm>:
-  FR<0x00, func, (outs CPU64Regs:$dst), (ins),
-     !strconcat(instr_asm, "\t$dst"), [], IIHiLo>;
+  FR<0x00, func, (outs CPU64Regs:$rd), (ins),
+     !strconcat(instr_asm, "\t$rd"), [], IIHiLo>;
 
 let rt = 0, rd = 0 in
 class MoveToLOHI64<bits<6> func, string instr_asm>:
-  FR<0x00, func, (outs), (ins CPU64Regs:$src),
-     !strconcat(instr_asm, "\t$src"), [], IIHiLo>;
+  FR<0x00, func, (outs), (ins CPU64Regs:$rs),
+     !strconcat(instr_asm, "\t$rs"), [], IIHiLo>;
 }
 
 // Count Leading Ones/Zeros in Word
 class CountLeading64<bits<6> func, string instr_asm, list<dag> pattern>:
-  FR<0x1c, func, (outs CPU64Regs:$dst), (ins CPU64Regs:$src),
-     !strconcat(instr_asm, "\t$dst, $src"), pattern, IIAlu>,
+  FR<0x1c, func, (outs CPU64Regs:$rd), (ins CPU64Regs:$rs),
+     !strconcat(instr_asm, "\t$rd, $rs"), pattern, IIAlu>,
      Requires<[HasBitCount]> {
   let shamt = 0;
   let rt = rd;
@@ -180,9 +180,9 @@ let Uses = [LO64] in
 
 /// Count Leading
 def DCLZ : CountLeading64<0x24, "dclz",
-                          [(set CPU64Regs:$dst, (ctlz CPU64Regs:$src))]>;
+                          [(set CPU64Regs:$rd, (ctlz CPU64Regs:$rs))]>;
 def DCLO : CountLeading64<0x25, "dclo",
-                          [(set CPU64Regs:$dst, (ctlz (not CPU64Regs:$src)))]>;
+                          [(set CPU64Regs:$rd, (ctlz (not CPU64Regs:$rs)))]>;
 
 //===----------------------------------------------------------------------===//
 //  Arbitrary patterns that map to one or more instructions
diff --git a/lib/Target/Mips/MipsCodeEmitter.cpp b/lib/Target/Mips/MipsCodeEmitter.cpp
index 9220d9c..23fabe3 100644
--- a/lib/Target/Mips/MipsCodeEmitter.cpp
+++ b/lib/Target/Mips/MipsCodeEmitter.cpp
@@ -105,6 +105,9 @@ class MipsCodeEmitter : public MachineFunctionPass {
     unsigned getRelocation(const MachineInstr &MI,
                            const MachineOperand &MO) const;
 
+    unsigned getMemEncoding(const MachineInstr &MI, unsigned OpNo) const;
+    unsigned getSizeExtEncoding(const MachineInstr &MI, unsigned OpNo) const;
+    unsigned getSizeInsEncoding(const MachineInstr &MI, unsigned OpNo) const;
   };
 }
 
@@ -153,6 +156,28 @@ unsigned MipsCodeEmitter::getRelocation(const MachineInstr &MI,
   return Mips::reloc_mips_lo;
 }
 
+unsigned MipsCodeEmitter::getMemEncoding(const MachineInstr &MI,
+                                          unsigned OpNo) const {
+  // Base register is encoded in bits 20-16, offset is encoded in bits 15-0.
+  assert(MI.getOperand(OpNo).isReg());
+  unsigned RegBits = getMachineOpValue(MI, MI.getOperand(OpNo)) << 16;
+  return
+    (getMachineOpValue(MI, MI.getOperand(OpNo+1)) & 0xFFFF) | RegBits;
+}
+
+unsigned MipsCodeEmitter::getSizeExtEncoding(const MachineInstr &MI,
+                                          unsigned OpNo) const {
+  // size is encoded as size-1.
+  return getMachineOpValue(MI, MI.getOperand(OpNo)) - 1;
+}
+
+unsigned MipsCodeEmitter::getSizeInsEncoding(const MachineInstr &MI,
+                                          unsigned OpNo) const {
+  // size is encoded as pos+size-1.
+  return getMachineOpValue(MI, MI.getOperand(OpNo-1)) +
+         getMachineOpValue(MI, MI.getOperand(OpNo)) - 1;
+}
+
 /// getMachineOpValue - Return binary encoding of operand. If the machine
 /// operand requires relocation, record the relocation and return zero.
 unsigned MipsCodeEmitter::getMachineOpValue(const MachineInstr &MI,
@@ -238,8 +263,4 @@ FunctionPass *llvm::createMipsJITCodeEmitterPass(MipsTargetMachine &TM,
   return new MipsCodeEmitter(TM, JCE);
 }
 
-unsigned MipsCodeEmitter::getBinaryCodeForInstr(const MachineInstr &MI) const {
- // this function will be automatically generated by the CodeEmitterGenerator
- // using TableGen
- return 0;
-}
+#include "MipsGenCodeEmitter.inc"
diff --git a/lib/Target/Mips/MipsInstrFPU.td b/lib/Target/Mips/MipsInstrFPU.td
index 2fb9d18..1fb779d 100644
--- a/lib/Target/Mips/MipsInstrFPU.td
+++ b/lib/Target/Mips/MipsInstrFPU.td
@@ -76,14 +76,16 @@ def IsNotSingleFloat : Predicate<"!Subtarget.isSingleFloat()">;
 // FP load.
 class FPLoad<bits<6> op, string opstr, PatFrag FOp, RegisterClass RC,
              Operand MemOpnd>:
-  FFI<op, (outs RC:$ft), (ins MemOpnd:$base),
-      !strconcat(opstr, "\t$ft, $base"), [(set RC:$ft, (FOp addr:$base))]>;
+  FMem<op, (outs RC:$ft), (ins MemOpnd:$addr),
+      !strconcat(opstr, "\t$ft, $addr"), [(set RC:$ft, (FOp addr:$addr))],
+      IILoad>;
 
 // FP store.
 class FPStore<bits<6> op, string opstr, PatFrag FOp, RegisterClass RC,
               Operand MemOpnd>:
-  FFI<op, (outs), (ins RC:$ft, MemOpnd:$base),
-      !strconcat(opstr, "\t$ft, $base"), [(store RC:$ft, addr:$base)]>;
+  FMem<op, (outs), (ins RC:$ft, MemOpnd:$addr),
+      !strconcat(opstr, "\t$ft, $addr"), [(store RC:$ft, addr:$addr)],
+      IIStore>;
 
 // Instructions that convert an FP value to 32-bit fixed point.
 multiclass FFR1_W_M<bits<6> funct, string opstr> {
@@ -158,22 +160,28 @@ defm FSQRT   : FFR1P_M<0x4, "sqrt", fsqrt>;
 // stores, and moves between floating-point and integer registers.
 // When defining instructions, we reference all 32-bit registers,
 // regardless of register aliasing.
-let fd = 0 in {
-  /// Move Control Registers From/To CPU Registers
-  def CFC1  : FFR<0x11, 0x0, 0x2, (outs CPURegs:$rt), (ins CCR:$fs),
+
+class FFRGPR<bits<5> _fmt, dag outs, dag ins, string asmstr, list<dag> pattern>:
+             FFR<0x11, 0x0, _fmt, outs, ins, asmstr, pattern> {
+  bits<5> rt;
+  let ft = rt;
+  let fd = 0;
+}
+
+/// Move Control Registers From/To CPU Registers
+def CFC1  : FFRGPR<0x2, (outs CPURegs:$rt), (ins CCR:$fs),
                   "cfc1\t$rt, $fs", []>;
 
-  def CTC1  : FFR<0x11, 0x0, 0x6, (outs CCR:$rt), (ins CPURegs:$fs),
-                  "ctc1\t$fs, $rt", []>;
+def CTC1  : FFRGPR<0x6, (outs CCR:$fs), (ins CPURegs:$rt),
+                  "ctc1\t$rt, $fs", []>;
 
-  def MFC1  : FFR<0x11, 0x00, 0x00, (outs CPURegs:$rt), (ins FGR32:$fs),
+def MFC1  : FFRGPR<0x00, (outs CPURegs:$rt), (ins FGR32:$fs),
                   "mfc1\t$rt, $fs",
                   [(set CPURegs:$rt, (bitconvert FGR32:$fs))]>;
 
-  def MTC1  : FFR<0x11, 0x00, 0x04, (outs FGR32:$fs), (ins CPURegs:$rt),
+def MTC1  : FFRGPR<0x04, (outs FGR32:$fs), (ins CPURegs:$rt),
                   "mtc1\t$rt, $fs",
                   [(set FGR32:$fs, (bitconvert CPURegs:$rt))]>;
-}
 
 def FMOV_S   : FFR1<0x6, 16, "mov", "s", FGR32, FGR32>;
 def FMOV_D32 : FFR1<0x6, 17, "mov", "d", AFGR64, AFGR64>,
@@ -203,7 +211,7 @@ let Predicates = [NotN64] in {
 }
 
 /// Floating-point Aritmetic
-defm FADD : FFR2P_M<0x10, "add", fadd, 1>;
+defm FADD : FFR2P_M<0x00, "add", fadd, 1>;
 defm FDIV : FFR2P_M<0x03, "div", fdiv>;
 defm FMUL : FFR2P_M<0x02, "mul", fmul, 1>;
 defm FSUB : FFR2P_M<0x01, "sub", fsub>;
@@ -218,12 +226,16 @@ def MIPS_BRANCH_T  : PatLeaf<(i32 1)>;
 
 /// Floating Point Branch of False/True (Likely)
 let isBranch=1, isTerminator=1, hasDelaySlot=1, base=0x8, Uses=[FCR31] in
-  class FBRANCH<PatLeaf op, string asmstr> : FFI<0x11, (outs),
-        (ins brtarget:$dst), !strconcat(asmstr, "\t$dst"),
-        [(MipsFPBrcond op, bb:$dst)]>;
+  class FBRANCH<bits<1> nd, bits<1> tf, PatLeaf op, string asmstr> :
+      FFI<0x11, (outs), (ins brtarget:$dst), !strconcat(asmstr, "\t$dst"),
+        [(MipsFPBrcond op, bb:$dst)]> {
+  let Inst{20-18} = 0;
+  let Inst{17} = nd;
+  let Inst{16} = tf;
+}
 
-def BC1F  : FBRANCH<MIPS_BRANCH_F,  "bc1f">;
-def BC1T  : FBRANCH<MIPS_BRANCH_T,  "bc1t">;
+def BC1F  : FBRANCH<0, 0, MIPS_BRANCH_F,  "bc1f">;
+def BC1T  : FBRANCH<0, 1, MIPS_BRANCH_T,  "bc1t">;
 
 //===----------------------------------------------------------------------===//
 // Floating Point Flag Conditions
@@ -249,11 +261,11 @@ def MIPS_FCOND_NGT  : PatLeaf<(i32 15)>;
 
 /// Floating Point Compare
 let Defs=[FCR31] in {
-  def FCMP_S32 : FCC<0x0, (outs), (ins FGR32:$fs, FGR32:$ft, condcode:$cc),
+  def FCMP_S32 : FCC<0x10, (outs), (ins FGR32:$fs, FGR32:$ft, condcode:$cc),
                      "c.$cc.s\t$fs, $ft",
                      [(MipsFPCmp FGR32:$fs, FGR32:$ft, imm:$cc)]>;
 
-  def FCMP_D32 : FCC<0x1, (outs), (ins AFGR64:$fs, AFGR64:$ft, condcode:$cc),
+  def FCMP_D32 : FCC<0x11, (outs), (ins AFGR64:$fs, AFGR64:$ft, condcode:$cc),
                      "c.$cc.d\t$fs, $ft",
                      [(MipsFPCmp AFGR64:$fs, AFGR64:$ft, imm:$cc)]>,
                      Requires<[NotFP64bit]>;
@@ -287,7 +299,8 @@ let Predicates = [NotFP64bit] in {
   defm : MovnPats<AFGR64, MOVN_D>;
 }
 
-let usesCustomInserter = 1, Uses = [FCR31], Constraints = "$F = $dst" in {
+let cc = 0, usesCustomInserter = 1, Uses = [FCR31],
+    Constraints = "$F = $dst" in {
 // flag:float, data:int
 class CondMovFPInt<SDNode cmov, bits<1> tf, string instr_asm> :
   FCMOV<tf, (outs CPURegs:$dst), (ins CPURegs:$T, CPURegs:$F),
@@ -295,6 +308,7 @@ class CondMovFPInt<SDNode cmov, bits<1> tf, string instr_asm> :
         [(set CPURegs:$dst, (cmov CPURegs:$T, CPURegs:$F))]>;
 
 // flag:float, data:float
+let cc = 0 in
 class CondMovFPFP<RegisterClass RC, SDNode cmov, bits<5> fmt, bits<1> tf,
                   string instr_asm> :
   FFCMOV<fmt, tf, (outs RC:$dst), (ins RC:$T, RC:$F),
diff --git a/lib/Target/Mips/MipsInstrFormats.td b/lib/Target/Mips/MipsInstrFormats.td
index d246a26..e1725fa 100644
--- a/lib/Target/Mips/MipsInstrFormats.td
+++ b/lib/Target/Mips/MipsInstrFormats.td
@@ -21,30 +21,55 @@
 //
 //===----------------------------------------------------------------------===//
 
+// Format specifies the encoding used by the instruction.  This is part of the
+// ad-hoc solution used to emit machine instruction encodings by our machine
+// code emitter.
+class Format<bits<4> val> {
+  bits<4> Value = val;
+}
+
+def Pseudo    : Format<0>;
+def FrmR      : Format<1>;
+def FrmI      : Format<2>;
+def FrmJ      : Format<3>;
+def FrmFR     : Format<4>;
+def FrmFI     : Format<5>;
+def FrmOther  : Format<6>; // Instruction w/ a custom format
+
 // Generic Mips Format
 class MipsInst<dag outs, dag ins, string asmstr, list<dag> pattern,
-               InstrItinClass itin>: Instruction
+               InstrItinClass itin, Format f>: Instruction
 {
   field bits<32> Inst;
+  Format Form = f;
 
   let Namespace = "Mips";
 
-  bits<6> opcode;
+  bits<6> Opcode = 0;
 
-  // Top 5 bits are the 'opcode' field
-  let Inst{31-26} = opcode;
+  // Top 6 bits are the 'opcode' field
+  let Inst{31-26} = Opcode;
 
-  dag OutOperandList = outs;
-  dag InOperandList  = ins;
+  let OutOperandList = outs;
+  let InOperandList  = ins;
 
   let AsmString   = asmstr;
   let Pattern     = pattern;
   let Itinerary   = itin;
+
+  //
+  // Attributes specific to Mips instructions...
+  //
+  bits<4> FormBits = Form.Value;
+
+  // TSFlags layout should be kept in sync with MipsInstrInfo.h.
+  let TSFlags{3-0}   = FormBits;
 }
 
 // Mips Pseudo Instructions Format
 class MipsPseudo<dag outs, dag ins, string asmstr, list<dag> pattern>:
-      MipsInst<outs, ins, asmstr, pattern, IIPseudo> {
+      MipsInst<outs, ins, asmstr, pattern, IIPseudo, Pseudo> {
+  let isCodeGenOnly = 1;
   let isPseudo = 1;
 }
 
@@ -54,7 +79,7 @@ class MipsPseudo<dag outs, dag ins, string asmstr, list<dag> pattern>:
 
 class FR<bits<6> op, bits<6> _funct, dag outs, dag ins, string asmstr,
          list<dag> pattern, InstrItinClass itin>:
-      MipsInst<outs, ins, asmstr, pattern, itin>
+      MipsInst<outs, ins, asmstr, pattern, itin, FrmR>
 {
   bits<5>  rd;
   bits<5>  rs;
@@ -62,7 +87,7 @@ class FR<bits<6> op, bits<6> _funct, dag outs, dag ins, string asmstr,
   bits<5>  shamt;
   bits<6>  funct;
 
-  let opcode = op;
+  let Opcode = op;
   let funct  = _funct;
 
   let Inst{25-21} = rs;
@@ -77,13 +102,13 @@ class FR<bits<6> op, bits<6> _funct, dag outs, dag ins, string asmstr,
 //===----------------------------------------------------------------------===//
 
 class FI<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern,
-         InstrItinClass itin>: MipsInst<outs, ins, asmstr, pattern, itin>
+         InstrItinClass itin>: MipsInst<outs, ins, asmstr, pattern, itin, FrmI>
 {
   bits<5>  rt;
   bits<5>  rs;
   bits<16> imm16;
 
-  let opcode = op;
+  let Opcode = op;
 
   let Inst{25-21} = rs;
   let Inst{20-16} = rt;
@@ -92,13 +117,13 @@ class FI<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern,
 
 class CBranchBase<bits<6> op, dag outs, dag ins, string asmstr,
                   list<dag> pattern, InstrItinClass itin>:
-  MipsInst<outs, ins, asmstr, pattern, itin>
+  MipsInst<outs, ins, asmstr, pattern, itin, FrmI>
 {
   bits<5>  rs;
   bits<5>  rt;
   bits<16> imm16;
 
-  let opcode = op;
+  let Opcode = op;
 
   let Inst{25-21} = rs;
   let Inst{20-16} = rt;
@@ -110,11 +135,11 @@ class CBranchBase<bits<6> op, dag outs, dag ins, string asmstr,
 //===----------------------------------------------------------------------===//
 
 class FJ<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern,
-         InstrItinClass itin>: MipsInst<outs, ins, asmstr, pattern, itin>
+         InstrItinClass itin>: MipsInst<outs, ins, asmstr, pattern, itin, FrmJ>
 {
   bits<26> addr;
 
-  let opcode = op;
+  let Opcode = op;
 
   let Inst{25-0} = addr;
 }
@@ -138,7 +163,7 @@ class FJ<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern,
 
 class FFR<bits<6> op, bits<6> _funct, bits<5> _fmt, dag outs, dag ins,
           string asmstr, list<dag> pattern> :
-          MipsInst<outs, ins, asmstr, pattern, NoItinerary>
+          MipsInst<outs, ins, asmstr, pattern, NoItinerary, FrmFR>
 {
   bits<5>  fd;
   bits<5>  fs;
@@ -146,7 +171,7 @@ class FFR<bits<6> op, bits<6> _funct, bits<5> _fmt, dag outs, dag ins,
   bits<5>  fmt;
   bits<6>  funct;
 
-  let opcode = op;
+  let Opcode = op;
   let funct  = _funct;
   let fmt    = _fmt;
 
@@ -162,13 +187,13 @@ class FFR<bits<6> op, bits<6> _funct, bits<5> _fmt, dag outs, dag ins,
 //===----------------------------------------------------------------------===//
 
 class FFI<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern>:
-          MipsInst<outs, ins, asmstr, pattern, NoItinerary>
+          MipsInst<outs, ins, asmstr, pattern, NoItinerary, FrmFI>
 {
   bits<5>  ft;
   bits<5>  base;
   bits<16> imm16;
 
-  let opcode = op;
+  let Opcode = op;
 
   let Inst{25-21} = base;
   let Inst{20-16} = ft;
@@ -180,14 +205,14 @@ class FFI<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern>:
 //===----------------------------------------------------------------------===//
 
 class FCC<bits<5> _fmt, dag outs, dag ins, string asmstr, list<dag> pattern> :
-          MipsInst<outs, ins, asmstr, pattern, NoItinerary>
+          MipsInst<outs, ins, asmstr, pattern, NoItinerary, FrmOther>
 {
   bits<5>  fs;
   bits<5>  ft;
   bits<4>  cc;
   bits<5>  fmt;
 
-  let opcode = 0x11;
+  let Opcode = 0x11;
   let fmt    = _fmt;
 
   let Inst{25-21} = fmt;
@@ -201,18 +226,18 @@ class FCC<bits<5> _fmt, dag outs, dag ins, string asmstr, list<dag> pattern> :
 
 class FCMOV<bits<1> _tf, dag outs, dag ins, string asmstr,
             list<dag> pattern> :
-  MipsInst<outs, ins, asmstr, pattern, NoItinerary>
+  MipsInst<outs, ins, asmstr, pattern, NoItinerary, FrmOther>
 {
   bits<5>  rd;
   bits<5>  rs;
-  bits<3>  N;
+  bits<3>  cc;
   bits<1>  tf;
 
-  let opcode = 0;
+  let Opcode = 0;
   let tf = _tf;
 
   let Inst{25-21} = rs;
-  let Inst{20-18} = N;
+  let Inst{20-18} = cc;
   let Inst{17} = 0;
   let Inst{16} = tf;
   let Inst{15-11} = rd;
@@ -222,20 +247,20 @@ class FCMOV<bits<1> _tf, dag outs, dag ins, string asmstr,
 
 class FFCMOV<bits<5> _fmt, bits<1> _tf, dag outs, dag ins, string asmstr,
              list<dag> pattern> :
-  MipsInst<outs, ins, asmstr, pattern, NoItinerary>
+  MipsInst<outs, ins, asmstr, pattern, NoItinerary, FrmOther>
 {
   bits<5>  fd;
   bits<5>  fs;
-  bits<3>  N;
+  bits<3>  cc;
   bits<5>  fmt;
   bits<1>  tf;
 
-  let opcode = 17;
+  let Opcode = 17;
   let fmt = _fmt;
   let tf = _tf;
 
   let Inst{25-21} = fmt;
-  let Inst{20-18} = N;
+  let Inst{20-18} = cc;
   let Inst{17} = 0;
   let Inst{16} = tf;
   let Inst{15-11} = fs;
diff --git a/lib/Target/Mips/MipsInstrInfo.td b/lib/Target/Mips/MipsInstrInfo.td
index 06b7de7..3fbd41e 100644
--- a/lib/Target/Mips/MipsInstrInfo.td
+++ b/lib/Target/Mips/MipsInstrInfo.td
@@ -153,6 +153,7 @@ def uimm16      : Operand<i32> {
 def mem : Operand<i32> {
   let PrintMethod = "printMemOperand";
   let MIOperandInfo = (ops CPURegs, simm16);
+  let EncoderMethod = "getMemEncoding";
 }
 
 def mem64 : Operand<i64> {
@@ -163,6 +164,17 @@ def mem64 : Operand<i64> {
 def mem_ea : Operand<i32> {
   let PrintMethod = "printMemOperandEA";
   let MIOperandInfo = (ops CPURegs, simm16);
+  let EncoderMethod = "getMemEncoding";
+}
+
+// size operand of ext instruction
+def size_ext : Operand<i32> {
+  let EncoderMethod = "getSizeExtEncoding";
+}
+
+// size operand of ins instruction
+def size_ins : Operand<i32> {
+  let EncoderMethod = "getSizeInsEncoding";
 }
 
 // Transformation Function - get the lower 16 bits.
@@ -271,14 +283,14 @@ class ArithOverflowR<bits<6> op, bits<6> func, string instr_asm,
 // Arithmetic and logical instructions with 2 register operands.
 class ArithLogicI<bits<6> op, string instr_asm, SDNode OpNode,
                   Operand Od, PatLeaf imm_type, RegisterClass RC> :
-  FI<op, (outs RC:$rt), (ins RC:$rs, Od:$i),
-     !strconcat(instr_asm, "\t$rt, $rs, $i"),
-     [(set RC:$rt, (OpNode RC:$rs, imm_type:$i))], IIAlu>;
+  FI<op, (outs RC:$rt), (ins RC:$rs, Od:$imm16),
+     !strconcat(instr_asm, "\t$rt, $rs, $imm16"),
+     [(set RC:$rt, (OpNode RC:$rs, imm_type:$imm16))], IIAlu>;
 
 class ArithOverflowI<bits<6> op, string instr_asm, SDNode OpNode,
                      Operand Od, PatLeaf imm_type, RegisterClass RC> :
-  FI<op, (outs RC:$rt), (ins RC:$rs, Od:$i),
-     !strconcat(instr_asm, "\t$rt, $rs, $i"), [], IIAlu>;
+  FI<op, (outs RC:$rt), (ins RC:$rs, Od:$imm16),
+     !strconcat(instr_asm, "\t$rt, $rs, $imm16"), [], IIAlu>;
 
 // Arithmetic Multiply ADD/SUB
 let rd = 0, shamt = 0, Defs = [HI, LO], Uses = [HI, LO] in
@@ -319,16 +331,23 @@ class LogicR_shift_rotate_reg<bits<6> func, bits<5> isRotate, string instr_asm,
 
 // Load Upper Imediate
 class LoadUpper<bits<6> op, string instr_asm>:
-  FI<op, (outs CPURegs:$rt), (ins uimm16:$imm),
-     !strconcat(instr_asm, "\t$rt, $imm"), [], IIAlu> {
+  FI<op, (outs CPURegs:$rt), (ins uimm16:$imm16),
+     !strconcat(instr_asm, "\t$rt, $imm16"), [], IIAlu> {
   let rs = 0;
 }
 
+class FMem<bits<6> op, dag outs, dag ins, string asmstr, list<dag> pattern,
+          InstrItinClass itin>: FFI<op, outs, ins, asmstr, pattern> {
+  bits<21> addr;
+  let Inst{25-21} = addr{20-16};
+  let Inst{15-0}  = addr{15-0};
+}
+
 // Memory Load/Store
 let canFoldAsLoad = 1 in
 class LoadM<bits<6> op, string instr_asm, PatFrag OpNode, RegisterClass RC,
             Operand MemOpnd, bit Pseudo>:
-  FI<op, (outs RC:$rt), (ins MemOpnd:$addr),
+  FMem<op, (outs RC:$rt), (ins MemOpnd:$addr),
      !strconcat(instr_asm, "\t$rt, $addr"),
      [(set RC:$rt, (OpNode addr:$addr))], IILoad> {
   let isPseudo = Pseudo;
@@ -336,7 +355,7 @@ class LoadM<bits<6> op, string instr_asm, PatFrag OpNode, RegisterClass RC,
 
 class StoreM<bits<6> op, string instr_asm, PatFrag OpNode, RegisterClass RC,
              Operand MemOpnd, bit Pseudo>:
-  FI<op, (outs), (ins RC:$rt, MemOpnd:$addr),
+  FMem<op, (outs), (ins RC:$rt, MemOpnd:$addr),
      !strconcat(instr_asm, "\t$rt, $addr"),
      [(OpNode RC:$rt, addr:$addr)], IIStore> {
   let isPseudo = Pseudo;
@@ -380,9 +399,9 @@ multiclass StoreM64<bits<6> op, string instr_asm, PatFrag OpNode,
 
 // Conditional Branch
 class CBranch<bits<6> op, string instr_asm, PatFrag cond_op, RegisterClass RC>:
-  CBranchBase<op, (outs), (ins RC:$rs, RC:$rt, brtarget:$offset),
-              !strconcat(instr_asm, "\t$rs, $rt, $offset"),
-              [(brcond (i32 (cond_op RC:$rs, RC:$rt)), bb:$offset)], IIBranch> {
+  CBranchBase<op, (outs), (ins RC:$rs, RC:$rt, brtarget:$imm16),
+              !strconcat(instr_asm, "\t$rs, $rt, $imm16"),
+              [(brcond (i32 (cond_op RC:$rs, RC:$rt)), bb:$imm16)], IIBranch> {
   let isBranch = 1;
   let isTerminator = 1;
   let hasDelaySlot = 1;
@@ -390,9 +409,9 @@ class CBranch<bits<6> op, string instr_asm, PatFrag cond_op, RegisterClass RC>:
 
 class CBranchZero<bits<6> op, bits<5> _rt, string instr_asm, PatFrag cond_op,
                   RegisterClass RC>:
-  CBranchBase<op, (outs), (ins RC:$rs, brtarget:$offset),
-              !strconcat(instr_asm, "\t$rs, $offset"),
-              [(brcond (i32 (cond_op RC:$rs, 0)), bb:$offset)], IIBranch> {
+  CBranchBase<op, (outs), (ins RC:$rs, brtarget:$imm16),
+              !strconcat(instr_asm, "\t$rs, $imm16"),
+              [(brcond (i32 (cond_op RC:$rs, 0)), bb:$imm16)], IIBranch> {
   let rt = _rt;
   let isBranch = 1;
   let isTerminator = 1;
@@ -411,9 +430,9 @@ class SetCC_R<bits<6> op, bits<6> func, string instr_asm, PatFrag cond_op,
 
 class SetCC_I<bits<6> op, string instr_asm, PatFrag cond_op, Operand Od,
               PatLeaf imm_type, RegisterClass RC>:
-  FI<op, (outs CPURegs:$rd), (ins RC:$rs, Od:$i),
-     !strconcat(instr_asm, "\t$rd, $rs, $i"),
-     [(set CPURegs:$rd, (cond_op RC:$rs, imm_type:$i))],
+  FI<op, (outs CPURegs:$rt), (ins RC:$rs, Od:$imm16),
+     !strconcat(instr_asm, "\t$rt, $rs, $imm16"),
+     [(set CPURegs:$rt, (cond_op RC:$rs, imm_type:$imm16))],
      IIAlu>;
 
 // Unconditional branch
@@ -450,10 +469,8 @@ let isCall=1, hasDelaySlot=1,
   }
 
   class BranchLink<string instr_asm>:
-    FI<0x1, (outs), (ins CPURegs:$rs, brtarget:$target, variable_ops),
-       !strconcat(instr_asm, "\t$rs, $target"), [], IIBranch> {
-    let rt = 0;
-  }
+    FI<0x1, (outs), (ins CPURegs:$rs, brtarget:$imm16, variable_ops),
+       !strconcat(instr_asm, "\t$rs, $imm16"), [], IIBranch>;
 }
 
 // Mul, Div
@@ -493,7 +510,7 @@ class MoveToLOHI<bits<6> func, string instr_asm>:
 }
 
 class EffectiveAddress<string instr_asm> :
-  FI<0x09, (outs CPURegs:$rt), (ins mem_ea:$addr),
+  FMem<0x09, (outs CPURegs:$rt), (ins mem_ea:$addr),
      instr_asm, [(set CPURegs:$rt, addr:$addr)], IIAlu>;
 
 // Count Leading Ones/Zeros in Word
@@ -507,7 +524,7 @@ class CountLeading<bits<6> func, string instr_asm, list<dag> pattern>:
 
 // Sign Extend in Register.
 class SignExtInReg<bits<5> sa, string instr_asm, ValueType vt>:
-  FR<0x3f, 0x20, (outs CPURegs:$rd), (ins CPURegs:$rt),
+  FR<0x1f, 0x20, (outs CPURegs:$rd), (ins CPURegs:$rt),
      !strconcat(instr_asm, "\t$rd, $rt"),
      [(set CPURegs:$rd, (sext_inreg CPURegs:$rt, vt))], NoItinerary> {
   let rs = 0;
@@ -685,20 +702,22 @@ defm USW     : StoreM32<0x2b, "usw", store_u, 1>;
 
 let hasSideEffects = 1 in
 def SYNC : MipsInst<(outs), (ins i32imm:$stype), "sync $stype",
-                    [(MipsSync imm:$stype)], NoItinerary>
+                    [(MipsSync imm:$stype)], NoItinerary, FrmOther>
 {
-  let opcode = 0;
+  bits<5> stype;
+  let Opcode = 0;
   let Inst{25-11} = 0;
+  let Inst{10-6} = stype;
   let Inst{5-0} = 15;
 }
 
 /// Load-linked, Store-conditional
 let mayLoad = 1 in
-  def LL    : FI<0x30, (outs CPURegs:$dst), (ins mem:$addr),
-              "ll\t$dst, $addr", [], IILoad>;
-let mayStore = 1, Constraints = "$src = $dst" in
-  def SC    : FI<0x38, (outs CPURegs:$dst), (ins CPURegs:$src, mem:$addr),
-              "sc\t$src, $addr", [], IIStore>;
+  def LL    : FMem<0x30, (outs CPURegs:$rt), (ins mem:$addr),
+              "ll\t$rt, $addr", [], IILoad>;
+let mayStore = 1, Constraints = "$rt = $dst" in
+  def SC    : FMem<0x38, (outs CPURegs:$dst), (ins CPURegs:$rt, mem:$addr),
+              "sc\t$rt, $addr", [], IIStore>;
 
 /// Jump and Branch Instructions
 def J       : JumpFJ<0x02, "j">;
@@ -710,15 +729,17 @@ def BEQ     : CBranch<0x04, "beq", seteq, CPURegs>;
 def BNE     : CBranch<0x05, "bne", setne, CPURegs>;
 def BGEZ    : CBranchZero<0x01, 1, "bgez", setge, CPURegs>;
 def BGTZ    : CBranchZero<0x07, 0, "bgtz", setgt, CPURegs>;
-def BLEZ    : CBranchZero<0x07, 0, "blez", setle, CPURegs>;
+def BLEZ    : CBranchZero<0x06, 0, "blez", setle, CPURegs>;
 def BLTZ    : CBranchZero<0x01, 0, "bltz", setlt, CPURegs>;
 
-def BGEZAL  : BranchLink<"bgezal">;
-def BLTZAL  : BranchLink<"bltzal">;
+let rt=0x11 in
+  def BGEZAL  : BranchLink<"bgezal">;
+let rt=0x10 in
+  def BLTZAL  : BranchLink<"bltzal">;
 
 let isReturn=1, isTerminator=1, hasDelaySlot=1,
-    isBarrier=1, hasCtrlDep=1, rs=0, rt=0, shamt=0 in
-  def RET : FR <0x00, 0x02, (outs), (ins CPURegs:$target),
+    isBarrier=1, hasCtrlDep=1, rd=0, rt=0, shamt=0 in
+  def RET : FR <0x00, 0x08, (outs), (ins CPURegs:$target),
                 "jr\t$target", [(MipsRet CPURegs:$target)], IIBranch>;
 
 /// Multiply and Divide Instructions.
@@ -797,14 +818,14 @@ def MUL   : ArithLogicR<0x1c, 0x02, "mul", mul, IIImul, CPURegs, 1>,
 def RDHWR : ReadHardware;
 
 def EXT : ExtIns<0, "ext", (outs CPURegs:$rt),
-                 (ins CPURegs:$rs, uimm16:$pos, uimm16:$sz),
+                 (ins CPURegs:$rs, uimm16:$pos, size_ext:$sz),
                  [(set CPURegs:$rt,
                    (MipsExt CPURegs:$rs, immZExt5:$pos, immZExt5:$sz))],
                  NoItinerary>;
 
 let Constraints = "$src = $rt" in
 def INS : ExtIns<4, "ins", (outs CPURegs:$rt),
-                 (ins CPURegs:$rs, uimm16:$pos, uimm16:$sz, CPURegs:$src),
+                 (ins CPURegs:$rs, uimm16:$pos, size_ins:$sz, CPURegs:$src),
                  [(set CPURegs:$rt,
                    (MipsIns CPURegs:$rs, immZExt5:$pos, immZExt5:$sz,
                     CPURegs:$src))],
diff --git a/lib/Target/Mips/MipsJITInfo.cpp b/lib/Target/Mips/MipsJITInfo.cpp
index 28c2b48..e3f6a75 100644
--- a/lib/Target/Mips/MipsJITInfo.cpp
+++ b/lib/Target/Mips/MipsJITInfo.cpp
@@ -57,11 +57,11 @@ void MipsCompilationCallback();
     ".globl " ASMPREFIX "MipsCompilationCallback\n"
     ASMPREFIX "MipsCompilationCallback:\n"
     ".ent " ASMPREFIX "MipsCompilationCallback\n"
-    ".frame  $29, 32, $31\n"
+    ".frame  $sp, 32, $ra\n"
     ".set  noreorder\n"
     ".cpload $t9\n"
 
-    "addiu $sp, $sp, -60\n"
+    "addiu $sp, $sp, -64\n"
     ".cprestore 16\n"
 
     // Save argument registers a0, a1, a2, a3, f12, f14 since they may contain
@@ -76,8 +76,8 @@ void MipsCompilationCallback();
     "sw $a3, 32($sp)\n"
     "sw $ra, 36($sp)\n"
     "sw $t8, 40($sp)\n"
-    "sdc1 $f12, 44($sp)\n"
-    "sdc1 $f14, 52($sp)\n"
+    "sdc1 $f12, 48($sp)\n"
+    "sdc1 $f14, 56($sp)\n"
 
     // t8 points at the end of function stub. Pass the beginning of the stub
     // to the MipsCompilationCallbackC.
@@ -92,9 +92,9 @@ void MipsCompilationCallback();
     "lw $a3, 32($sp)\n"
     "lw $ra, 36($sp)\n"
     "lw $t8, 40($sp)\n"
-    "ldc1 $f12, 44($sp)\n"
-    "ldc1 $f14, 52($sp)\n"
-    "addiu $sp, $sp, 60\n"
+    "ldc1 $f12, 48($sp)\n"
+    "ldc1 $f14, 56($sp)\n"
+    "addiu $sp, $sp, 64\n"
 
     // Jump to the (newly modified) stub to invoke the real function.
     "addiu $t8, $t8, -16\n"
diff --git a/lib/Target/PowerPC/PPCFrameLowering.cpp b/lib/Target/PowerPC/PPCFrameLowering.cpp
index 7dead10..0b85fea 100644
--- a/lib/Target/PowerPC/PPCFrameLowering.cpp
+++ b/lib/Target/PowerPC/PPCFrameLowering.cpp
@@ -490,10 +490,8 @@ void PPCFrameLowering::emitPrologue(MachineFunction &MF) const {
 
       // This is a bit of a hack: CR2LT, CR2GT, CR2EQ and CR2UN are just
       // subregisters of CR2. We just need to emit a move of CR2.
-      if (Reg == PPC::CR2LT || Reg == PPC::CR2GT || Reg == PPC::CR2EQ)
+      if (PPC::CRBITRCRegisterClass->contains(Reg))
         continue;
-      if (Reg == PPC::CR2UN)
-        Reg = PPC::CR2;
 
       MachineLocation CSDst(MachineLocation::VirtualFP, Offset);
       MachineLocation CSSrc(Reg);
diff --git a/lib/Target/X86/X86CodeEmitter.cpp b/lib/Target/X86/X86CodeEmitter.cpp
index aeff03a..f939510 100644
--- a/lib/Target/X86/X86CodeEmitter.cpp
+++ b/lib/Target/X86/X86CodeEmitter.cpp
@@ -589,6 +589,13 @@ void Emitter<CodeEmitter>::emitMemModRMByte(const MachineInstr &MI,
   }
 }
 
+static const MCInstrDesc *UpdateOp(MachineInstr &MI, const X86InstrInfo *II,
+                                   unsigned Opcode) {
+  const MCInstrDesc *Desc = &II->get(Opcode);
+  MI.setDesc(*Desc);
+  return Desc;
+}
+
 template<class CodeEmitter>
 void Emitter<CodeEmitter>::emitInstruction(MachineInstr &MI,
                                            const MCInstrDesc *Desc) {
@@ -596,15 +603,23 @@ void Emitter<CodeEmitter>::emitInstruction(MachineInstr &MI,
   
   // If this is a pseudo instruction, lower it.
   switch (Desc->getOpcode()) {
-  case X86::ADD16rr_DB:   Desc = &II->get(X86::OR16rr); MI.setDesc(*Desc);break;
-  case X86::ADD32rr_DB:   Desc = &II->get(X86::OR32rr); MI.setDesc(*Desc);break;
-  case X86::ADD64rr_DB:   Desc = &II->get(X86::OR64rr); MI.setDesc(*Desc);break;
-  case X86::ADD16ri_DB:   Desc = &II->get(X86::OR16ri); MI.setDesc(*Desc);break;
-  case X86::ADD32ri_DB:   Desc = &II->get(X86::OR32ri); MI.setDesc(*Desc);break;
-  case X86::ADD64ri32_DB:Desc = &II->get(X86::OR64ri32);MI.setDesc(*Desc);break;
-  case X86::ADD16ri8_DB:  Desc = &II->get(X86::OR16ri8);MI.setDesc(*Desc);break;
-  case X86::ADD32ri8_DB:  Desc = &II->get(X86::OR32ri8);MI.setDesc(*Desc);break;
-  case X86::ADD64ri8_DB:  Desc = &II->get(X86::OR64ri8);MI.setDesc(*Desc);break;
+  case X86::ADD16rr_DB:      Desc = UpdateOp(MI, II, X86::OR16rr); break;
+  case X86::ADD32rr_DB:      Desc = UpdateOp(MI, II, X86::OR32rr); break;
+  case X86::ADD64rr_DB:      Desc = UpdateOp(MI, II, X86::OR64rr); break;
+  case X86::ADD16ri_DB:      Desc = UpdateOp(MI, II, X86::OR16ri); break;
+  case X86::ADD32ri_DB:      Desc = UpdateOp(MI, II, X86::OR32ri); break;
+  case X86::ADD64ri32_DB:    Desc = UpdateOp(MI, II, X86::OR64ri32); break;
+  case X86::ADD16ri8_DB:     Desc = UpdateOp(MI, II, X86::OR16ri8); break;
+  case X86::ADD32ri8_DB:     Desc = UpdateOp(MI, II, X86::OR32ri8); break;
+  case X86::ADD64ri8_DB:     Desc = UpdateOp(MI, II, X86::OR64ri8); break;
+  case X86::ACQUIRE_MOV8rm:  Desc = UpdateOp(MI, II, X86::MOV8rm); break;
+  case X86::ACQUIRE_MOV16rm: Desc = UpdateOp(MI, II, X86::MOV16rm); break;
+  case X86::ACQUIRE_MOV32rm: Desc = UpdateOp(MI, II, X86::MOV32rm); break;
+  case X86::ACQUIRE_MOV64rm: Desc = UpdateOp(MI, II, X86::MOV64rm); break;
+  case X86::RELEASE_MOV8mr:  Desc = UpdateOp(MI, II, X86::MOV8mr); break;
+  case X86::RELEASE_MOV16mr: Desc = UpdateOp(MI, II, X86::MOV16mr); break;
+  case X86::RELEASE_MOV32mr: Desc = UpdateOp(MI, II, X86::MOV32mr); break;
+  case X86::RELEASE_MOV64mr: Desc = UpdateOp(MI, II, X86::MOV64mr); break;
   }
   
 
diff --git a/lib/Transforms/InstCombine/InstructionCombining.cpp b/lib/Transforms/InstCombine/InstructionCombining.cpp
index 92874b9..c15b805 100644
--- a/lib/Transforms/InstCombine/InstructionCombining.cpp
+++ b/lib/Transforms/InstCombine/InstructionCombining.cpp
@@ -2025,9 +2025,10 @@ bool InstCombiner::DoOneIteration(Function &F, unsigned Iteration) {
         BasicBlock *InstParent = I->getParent();
         BasicBlock::iterator InsertPos = I;
 
-        if (!isa<PHINode>(Result))        // If combining a PHI, don't insert
-          while (isa<PHINode>(InsertPos)) // middle of a block of PHIs.
-            ++InsertPos;
+        // If we replace a PHI with something that isn't a PHI, fix up the
+        // insertion point.
+        if (!isa<PHINode>(Result) && isa<PHINode>(InsertPos))
+          InsertPos = InstParent->getFirstInsertionPt();
 
         InstParent->getInstList().insert(InsertPos, Result);