diff --git a/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp b/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
index c5288c4..42e7e22 100644
--- a/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
+++ b/lib/CodeGen/SelectionDAG/LegalizeDAG.cpp
@@ -622,6 +622,7 @@
   case ISD::TargetConstantFP:
   case ISD::TargetConstantPool:
   case ISD::TargetGlobalAddress:
+  case ISD::TargetGlobalTLSAddress:
   case ISD::TargetExternalSymbol:
   case ISD::VALUETYPE:
   case ISD::SRCVALUE:
@@ -653,6 +654,7 @@
     assert(0 && "Do not know how to legalize this operator!");
     abort();
   case ISD::GlobalAddress:
+  case ISD::GlobalTLSAddress:
   case ISD::ExternalSymbol:
   case ISD::ConstantPool:
   case ISD::JumpTable: // Nothing to do.
diff --git a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
index 7c8a94e..75fbf8a 100644
--- a/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
+++ b/lib/CodeGen/SelectionDAG/SelectionDAG.cpp
@@ -14,6 +14,7 @@
 #include "llvm/CodeGen/SelectionDAG.h"
 #include "llvm/Constants.h"
 #include "llvm/GlobalValue.h"
+#include "llvm/GlobalVariable.h"
 #include "llvm/Intrinsics.h"
 #include "llvm/Assembly/Writer.h"
 #include "llvm/CodeGen/MachineBasicBlock.h"
@@ -296,7 +297,9 @@
     ID.AddDouble(cast<ConstantFPSDNode>(N)->getValue());
     break;
   case ISD::TargetGlobalAddress:
-  case ISD::GlobalAddress: {
+  case ISD::GlobalAddress:
+  case ISD::TargetGlobalTLSAddress:
+  case ISD::GlobalTLSAddress: {
     GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(N);
     ID.AddPointer(GA->getGlobal());
     ID.AddInteger(GA->getOffset());
@@ -692,7 +695,12 @@
 SDOperand SelectionDAG::getGlobalAddress(const GlobalValue *GV,
                                          MVT::ValueType VT, int Offset,
                                          bool isTargetGA) {
-  unsigned Opc = isTargetGA ? ISD::TargetGlobalAddress : ISD::GlobalAddress;
+  const GlobalVariable *GVar = dyn_cast<GlobalVariable>(GV);
+  unsigned Opc;
+  if (GVar && GVar->isThreadLocal())
+    Opc = isTargetGA ? ISD::TargetGlobalTLSAddress : ISD::GlobalTLSAddress;
+  else
+    Opc = isTargetGA ? ISD::TargetGlobalAddress : ISD::GlobalAddress;
   FoldingSetNodeID ID;
   AddNodeIDNode(ID, Opc, getVTList(VT), 0, 0);
   ID.AddPointer(GV);
@@ -2282,6 +2290,14 @@
   SDOperand Ops[] = { Op1, Op2 };
   return getNode(ISD::BUILTIN_OP_END+Opcode, VTs, 3, Ops, 2).Val;
 }
+SDNode *SelectionDAG::getTargetNode(unsigned Opcode, MVT::ValueType VT1,
+                                    MVT::ValueType VT2, MVT::ValueType VT3,
+                                    SDOperand Op1, SDOperand Op2,
+                                    SDOperand Op3) {
+  const MVT::ValueType *VTs = getNodeValueTypes(VT1, VT2, VT3);
+  SDOperand Ops[] = { Op1, Op2, Op3 };
+  return getNode(ISD::BUILTIN_OP_END+Opcode, VTs, 3, Ops, 3).Val;
+}
 SDNode *SelectionDAG::getTargetNode(unsigned Opcode, MVT::ValueType VT1, 
                                     MVT::ValueType VT2, MVT::ValueType VT3,
                                     const SDOperand *Ops, unsigned NumOps) {
@@ -2702,6 +2718,7 @@
   case ISD::Constant:      return "Constant";
   case ISD::ConstantFP:    return "ConstantFP";
   case ISD::GlobalAddress: return "GlobalAddress";
+  case ISD::GlobalTLSAddress: return "GlobalTLSAddress";
   case ISD::FrameIndex:    return "FrameIndex";
   case ISD::JumpTable:     return "JumpTable";
   case ISD::GLOBAL_OFFSET_TABLE: return "GLOBAL_OFFSET_TABLE";
@@ -2725,6 +2742,7 @@
   case ISD::TargetConstant: return "TargetConstant";
   case ISD::TargetConstantFP:return "TargetConstantFP";
   case ISD::TargetGlobalAddress: return "TargetGlobalAddress";
+  case ISD::TargetGlobalTLSAddress: return "TargetGlobalTLSAddress";
   case ISD::TargetFrameIndex: return "TargetFrameIndex";
   case ISD::TargetJumpTable:  return "TargetJumpTable";
   case ISD::TargetConstantPool:  return "TargetConstantPool";
diff --git a/lib/Target/TargetAsmInfo.cpp b/lib/Target/TargetAsmInfo.cpp
index cf3f6ee..62c8b90 100644
--- a/lib/Target/TargetAsmInfo.cpp
+++ b/lib/Target/TargetAsmInfo.cpp
@@ -20,6 +20,8 @@
   TextSection(".text"),
   DataSection(".data"),
   BSSSection(".bss"),
+  TLSDataSection("\t.section .tdata,\"awT\",@progbits"),
+  TLSBSSSection("\t.section .tbss,\"awT\",@nobits"),
   ZeroFillDirective(0),
   AddressSize(4),
   NeedsSet(false),
diff --git a/lib/Target/TargetSelectionDAG.td b/lib/Target/TargetSelectionDAG.td
index 17c0ddd..3d728e9 100644
--- a/lib/Target/TargetSelectionDAG.td
+++ b/lib/Target/TargetSelectionDAG.td
@@ -211,6 +211,10 @@
                         "GlobalAddressSDNode">;
 def tglobaladdr : SDNode<"ISD::TargetGlobalAddress",  SDTPtrLeaf, [],
                          "GlobalAddressSDNode">;
+def globaltlsaddr : SDNode<"ISD::GlobalTLSAddress",         SDTPtrLeaf, [],
+                          "GlobalAddressSDNode">;
+def tglobaltlsaddr : SDNode<"ISD::TargetGlobalTLSAddress",  SDTPtrLeaf, [],
+                           "GlobalAddressSDNode">;
 def constpool   : SDNode<"ISD::ConstantPool",         SDTPtrLeaf, [],
                          "ConstantPoolSDNode">;
 def tconstpool  : SDNode<"ISD::TargetConstantPool",   SDTPtrLeaf, [],
diff --git a/lib/Target/X86/X86ATTAsmPrinter.cpp b/lib/Target/X86/X86ATTAsmPrinter.cpp
index f80e6e8..f36de92 100755
--- a/lib/Target/X86/X86ATTAsmPrinter.cpp
+++ b/lib/Target/X86/X86ATTAsmPrinter.cpp
@@ -272,9 +272,13 @@
   case MachineOperand::MO_GlobalAddress: {
     bool isCallOp = Modifier && !strcmp(Modifier, "call");
     bool isMemOp  = Modifier && !strcmp(Modifier, "mem");
-    if (!isMemOp && !isCallOp) O << '$';
 
     GlobalValue *GV = MO.getGlobal();
+    GlobalVariable *GVar = dyn_cast<GlobalVariable>(GV);
+    bool isThreadLocal = GVar && GVar->isThreadLocal();
+
+    if (!isMemOp && !isCallOp && !isThreadLocal) O << '$';
+
     std::string Name = Mang->getValueName(GV);
     X86SharedAsmPrinter::decorateName(Name, GV);
     
@@ -328,7 +332,15 @@
     else if (Offset < 0)
       O << Offset;
 
-    if (isMemOp) {
+    if (isThreadLocal) {
+      if (TM.getRelocationModel() == Reloc::PIC_)
+        O << "@TLSGD"; // general dynamic TLS model
+      else
+        if (GV->isDeclaration())
+          O << "@INDNTPOFF"; // initial exec TLS model
+        else
+          O << "@NTPOFF"; // local exec TLS model
+    } else if (isMemOp) {
       if (printGOT(TM, Subtarget)) {
         if (Subtarget->GVRequiresExtraLoad(GV, TM, false))
           O << "@GOT";
diff --git a/lib/Target/X86/X86AsmPrinter.cpp b/lib/Target/X86/X86AsmPrinter.cpp
index bc103d7..76037f1 100644
--- a/lib/Target/X86/X86AsmPrinter.cpp
+++ b/lib/Target/X86/X86AsmPrinter.cpp
@@ -171,7 +171,7 @@
         }
       }
       
-      if (!I->hasSection() &&
+      if (!I->hasSection() && !I->isThreadLocal() &&
           (I->hasInternalLinkage() || I->hasWeakLinkage() ||
            I->hasLinkOnceLinkage())) {
         if (Size == 0) Size = 1;   // .comm Foo, 0 is undefined, avoid it.
@@ -256,9 +256,13 @@
         SwitchToDataSection(SectionName.c_str());
       } else {
         if (C->isNullValue() && !NoZerosInBSS && TAI->getBSSSection())
-          SwitchToDataSection(TAI->getBSSSection(), I);
+          SwitchToDataSection(I->isThreadLocal() ? TAI->getTLSBSSSection() :
+                              TAI->getBSSSection(), I);
         else if (!I->isConstant())
-          SwitchToDataSection(TAI->getDataSection(), I);
+          SwitchToDataSection(I->isThreadLocal() ? TAI->getTLSDataSection() :
+                              TAI->getDataSection(), I);
+        else if (I->isThreadLocal())
+          SwitchToDataSection(TAI->getTLSDataSection());
         else {
           // Read-only data.
           bool HasReloc = C->ContainsRelocations();
diff --git a/lib/Target/X86/X86ISelLowering.cpp b/lib/Target/X86/X86ISelLowering.cpp
index 747a15b..b37e47c 100644
--- a/lib/Target/X86/X86ISelLowering.cpp
+++ b/lib/Target/X86/X86ISelLowering.cpp
@@ -20,6 +20,7 @@
 #include "llvm/CallingConv.h"
 #include "llvm/Constants.h"
 #include "llvm/DerivedTypes.h"
+#include "llvm/GlobalVariable.h"
 #include "llvm/Function.h"
 #include "llvm/Intrinsics.h"
 #include "llvm/ADT/VectorExtras.h"
@@ -200,6 +201,7 @@
   setOperationAction(ISD::ConstantPool    , MVT::i32  , Custom);
   setOperationAction(ISD::JumpTable       , MVT::i32  , Custom);
   setOperationAction(ISD::GlobalAddress   , MVT::i32  , Custom);
+  setOperationAction(ISD::GlobalTLSAddress, MVT::i32  , Custom);
   setOperationAction(ISD::ExternalSymbol  , MVT::i32  , Custom);
   if (Subtarget->is64Bit()) {
     setOperationAction(ISD::ConstantPool  , MVT::i64  , Custom);
@@ -2943,6 +2945,76 @@
   return Result;
 }
 
+// Lower ISD::GlobalTLSAddress using the "general dynamic" model
+static SDOperand
+LowerToTLSGeneralDynamicModel(GlobalAddressSDNode *GA, SelectionDAG &DAG,
+                              const MVT::ValueType PtrVT) {
+  SDOperand InFlag;
+  SDOperand Chain = DAG.getCopyToReg(DAG.getEntryNode(), X86::EBX,
+                                     DAG.getNode(X86ISD::GlobalBaseReg,
+                                                 PtrVT), InFlag);
+  InFlag = Chain.getValue(1);
+
+  // emit leal symbol@TLSGD(,%ebx,1), %eax
+  SDVTList NodeTys = DAG.getVTList(PtrVT, MVT::Other, MVT::Flag);
+  SDOperand TGA = DAG.getTargetGlobalAddress(GA->getGlobal(),
+                                             GA->getValueType(0),
+                                             GA->getOffset());
+  SDOperand Ops[] = { Chain,  TGA, InFlag };
+  SDOperand Result = DAG.getNode(X86ISD::TLSADDR, NodeTys, Ops, 3);
+  InFlag = Result.getValue(2);
+  Chain = Result.getValue(1);
+
+  // call ___tls_get_addr. This function receives its argument in
+  // the register EAX.
+  Chain = DAG.getCopyToReg(Chain, X86::EAX, Result, InFlag);
+  InFlag = Chain.getValue(1);
+
+  NodeTys = DAG.getVTList(MVT::Other, MVT::Flag);
+  SDOperand Ops1[] = { Chain,
+                      DAG.getTargetExternalSymbol("___tls_get_addr",
+                                                  PtrVT),
+                      DAG.getRegister(X86::EAX, PtrVT),
+                      DAG.getRegister(X86::EBX, PtrVT),
+                      InFlag };
+  Chain = DAG.getNode(X86ISD::CALL, NodeTys, Ops1, 5);
+  InFlag = Chain.getValue(1);
+
+  return DAG.getCopyFromReg(Chain, X86::EAX, PtrVT, InFlag);
+}
+
+// Lower ISD::GlobalTLSAddress using the "initial exec" (for no-pic) or
+// "local exec" model.
+static SDOperand
+LowerToTLSExecModel(GlobalAddressSDNode *GA, SelectionDAG &DAG,
+                         const MVT::ValueType PtrVT) {
+  // Get the Thread Pointer
+  SDOperand ThreadPointer = DAG.getNode(X86ISD::THREAD_POINTER, PtrVT);
+  // emit "addl x@ntpoff,%eax" (local exec) or "addl x@indntpoff,%eax" (initial
+  // exec)
+  SDOperand TGA = DAG.getTargetGlobalAddress(GA->getGlobal(),
+                                             GA->getValueType(0),
+                                             GA->getOffset());
+  SDOperand Offset = DAG.getNode(X86ISD::Wrapper, PtrVT, TGA);
+  // The address of the thread local variable is the add of the thread
+  // pointer with the offset of the variable.
+  return DAG.getNode(ISD::ADD, PtrVT, ThreadPointer, Offset);
+}
+
+SDOperand
+X86TargetLowering::LowerGlobalTLSAddress(SDOperand Op, SelectionDAG &DAG) {
+  // TODO: implement the "local dynamic" model
+  // TODO: implement the "initial exec"model for pic executables 
+  assert(!Subtarget->is64Bit() && "TLS not implemented for X86_64");
+  GlobalAddressSDNode *GA = cast<GlobalAddressSDNode>(Op);
+  // If the relocation model is PIC, use the "General Dynamic" TLS Model,
+  // otherwise use the "Local Exec"TLS Model
+  if (getTargetMachine().getRelocationModel() == Reloc::PIC_)
+    return LowerToTLSGeneralDynamicModel(GA, DAG, getPointerTy());
+  else
+    return LowerToTLSExecModel(GA, DAG, getPointerTy());
+}
+
 SDOperand
 X86TargetLowering::LowerExternalSymbol(SDOperand Op, SelectionDAG &DAG) {
   const char *Sym = cast<ExternalSymbolSDNode>(Op)->getSymbol();
@@ -4022,6 +4094,7 @@
   case ISD::SCALAR_TO_VECTOR:   return LowerSCALAR_TO_VECTOR(Op, DAG);
   case ISD::ConstantPool:       return LowerConstantPool(Op, DAG);
   case ISD::GlobalAddress:      return LowerGlobalAddress(Op, DAG);
+  case ISD::GlobalTLSAddress:   return LowerGlobalTLSAddress(Op, DAG);
   case ISD::ExternalSymbol:     return LowerExternalSymbol(Op, DAG);
   case ISD::SHL_PARTS:
   case ISD::SRA_PARTS:
@@ -4090,6 +4163,8 @@
   case X86ISD::PINSRW:             return "X86ISD::PINSRW";
   case X86ISD::FMAX:               return "X86ISD::FMAX";
   case X86ISD::FMIN:               return "X86ISD::FMIN";
+  case X86ISD::TLSADDR:            return "X86ISD::TLSADDR";
+  case X86ISD::THREAD_POINTER:     return "X86ISD::THREAD_POINTER";
   }
 }
 
diff --git a/lib/Target/X86/X86ISelLowering.h b/lib/Target/X86/X86ISelLowering.h
index 8b9c269..88064d3 100644
--- a/lib/Target/X86/X86ISelLowering.h
+++ b/lib/Target/X86/X86ISelLowering.h
@@ -176,7 +176,9 @@
 
       /// FMAX, FMIN - Floating point max and min.
       ///
-      FMAX, FMIN
+      FMAX, FMIN,
+      // Thread Local Storage
+      TLSADDR, THREAD_POINTER
     };
   }
 
@@ -386,6 +388,7 @@
     SDOperand LowerSCALAR_TO_VECTOR(SDOperand Op, SelectionDAG &DAG);
     SDOperand LowerConstantPool(SDOperand Op, SelectionDAG &DAG);
     SDOperand LowerGlobalAddress(SDOperand Op, SelectionDAG &DAG);
+    SDOperand LowerGlobalTLSAddress(SDOperand Op, SelectionDAG &DAG);
     SDOperand LowerExternalSymbol(SDOperand Op, SelectionDAG &DAG);
     SDOperand LowerShift(SDOperand Op, SelectionDAG &DAG);
     SDOperand LowerSINT_TO_FP(SDOperand Op, SelectionDAG &DAG);
diff --git a/lib/Target/X86/X86InstrInfo.td b/lib/Target/X86/X86InstrInfo.td
index bfdaff6..fdc2bc9 100644
--- a/lib/Target/X86/X86InstrInfo.td
+++ b/lib/Target/X86/X86InstrInfo.td
@@ -47,6 +47,10 @@
 
 def SDTX86Wrapper : SDTypeProfile<1, 1, [SDTCisSameAs<0, 1>, SDTCisPtrTy<0>]>;
 
+def SDT_X86TLSADDR : SDTypeProfile<1, 1, [SDTCisPtrTy<0>, SDTCisInt<1>]>;
+
+def SDT_X86TLSTP : SDTypeProfile<1, 0, [SDTCisPtrTy<0>]>;
+
 def X86shld    : SDNode<"X86ISD::SHLD",     SDTIntShiftDOp>;
 def X86shrd    : SDNode<"X86ISD::SHRD",     SDTIntShiftDOp>;
 
@@ -87,6 +91,11 @@
 def X86Wrapper    : SDNode<"X86ISD::Wrapper",     SDTX86Wrapper>;
 def X86WrapperRIP : SDNode<"X86ISD::WrapperRIP",  SDTX86Wrapper>;
 
+def X86tlsaddr : SDNode<"X86ISD::TLSADDR", SDT_X86TLSADDR,
+                        [SDNPHasChain, SDNPInFlag, SDNPOutFlag]>;
+def X86TLStp : SDNode<"X86ISD::THREAD_POINTER", SDT_X86TLSTP, []>;
+
+
 //===----------------------------------------------------------------------===//
 // X86 Operand Definitions.
 //
@@ -2449,6 +2458,19 @@
                 "mov{l} {$src, $dst|$dst, $src}", []>;
 
 //===----------------------------------------------------------------------===//
+// Thread Local Storage Instructions
+//
+
+def TLS_addr : I<0, Pseudo, (ops GR32:$dst, i32imm:$sym),
+               "leal $sym(,%ebx,1), $dst",
+               [(set GR32:$dst, (X86tlsaddr tglobaltlsaddr:$sym))]>,
+               Imp<[EBX],[]>;
+
+def TLS_tp : I<0, Pseudo, (ops GR32:$dst),
+               "movl %gs:0, $dst",
+               [(set GR32:$dst, X86TLStp)]>;
+
+//===----------------------------------------------------------------------===//
 // DWARF Pseudo Instructions
 //