diff --git a/lib/Target/X86/X86InstrInfo.cpp b/lib/Target/X86/X86InstrInfo.cpp
index 9d304b52dbd..a402a382c86 100644
--- a/lib/Target/X86/X86InstrInfo.cpp
+++ b/lib/Target/X86/X86InstrInfo.cpp
@@ -126,15 +126,24 @@ MachineInstr *X86InstrInfo::convertToThreeAddress(MachineInstr *MI) const {
 ///
 MachineInstr *X86InstrInfo::commuteInstruction(MachineInstr *MI) const {
   switch (MI->getOpcode()) {
+  case X86::SHRD16rri8: // A = SHRD16rri8 B, C, I -> A = SHLD16rri8 C, B, (16-I)
+  case X86::SHLD16rri8: // A = SHLD16rri8 B, C, I -> A = SHRD16rri8 C, B, (16-I)
   case X86::SHRD32rri8: // A = SHRD32rri8 B, C, I -> A = SHLD32rri8 C, B, (32-I)
   case X86::SHLD32rri8:{// A = SHLD32rri8 B, C, I -> A = SHRD32rri8 C, B, (32-I)
+    unsigned Opc;
+    unsigned Size;
+    switch (MI->getOpcode()) {
+    default: assert(0 && "Unreachable!");
+    case X86::SHRD16rri8: Size = 16; Opc = X86::SHLD16rri8; break;
+    case X86::SHLD16rri8: Size = 16; Opc = X86::SHRD16rri8; break;
+    case X86::SHRD32rri8: Size = 32; Opc = X86::SHLD32rri8; break;
+    case X86::SHLD32rri8: Size = 32; Opc = X86::SHRD32rri8; break;
+    }
     unsigned Amt = MI->getOperand(3).getImmedValue();
     unsigned A = MI->getOperand(0).getReg();
     unsigned B = MI->getOperand(1).getReg();
     unsigned C = MI->getOperand(2).getReg();
-    unsigned Opc = X86::SHRD32rri8;
-    if (MI->getOpcode() == X86::SHRD32rri8) Opc = X86::SHLD32rri8;
-    return BuildMI(Opc, 3, A).addReg(B).addReg(C).addImm(32-Amt);
+    return BuildMI(Opc, 3, A).addReg(B).addReg(C).addImm(Size-Amt);
   }
   default:
     return TargetInstrInfo::commuteInstruction(MI);
diff --git a/lib/Target/X86/X86InstrInfo.td b/lib/Target/X86/X86InstrInfo.td
index d50133bfdc0..7534393ed49 100644
--- a/lib/Target/X86/X86InstrInfo.td
+++ b/lib/Target/X86/X86InstrInfo.td
@@ -881,6 +881,12 @@ def SHLD32rrCL : I<0xA5, MRMDestReg, (ops R32:$dst, R32:$src1, R32:$src2),
 def SHRD32rrCL : I<0xAD, MRMDestReg, (ops R32:$dst, R32:$src1, R32:$src2),
                    "shrd{l} {%cl, $src2, $dst|$dst, $src2, %CL}">,
                    Imp<[CL],[]>, TB;
+def SHLD16rrCL : I<0xA5, MRMDestReg, (ops R16:$dst, R16:$src1, R16:$src2),
+                   "shld{w} {%cl, $src2, $dst|$dst, $src2, %CL}">,
+                   Imp<[CL],[]>, TB, OpSize;
+def SHRD16rrCL : I<0xAD, MRMDestReg, (ops R16:$dst, R16:$src1, R16:$src2),
+                   "shrd{w} {%cl, $src2, $dst|$dst, $src2, %CL}">,
+                   Imp<[CL],[]>, TB, OpSize;
 
 let isCommutable = 1 in {  // These instructions commute to each other.
 def SHLD32rri8 : Ii8<0xA4, MRMDestReg,
@@ -889,6 +895,14 @@ def SHLD32rri8 : Ii8<0xA4, MRMDestReg,
 def SHRD32rri8 : Ii8<0xAC, MRMDestReg,
                      (ops R32:$dst, R32:$src1, R32:$src2, i8imm:$src3),
                      "shrd{l} {$src3, $src2, $dst|$dst, $src2, $src3}">, TB;
+def SHLD16rri8 : Ii8<0xA4, MRMDestReg,
+                     (ops R16:$dst, R16:$src1, R16:$src2, i8imm:$src3),
+                     "shld{w} {$src3, $src2, $dst|$dst, $src2, $src3}">,
+                     TB, OpSize;
+def SHRD16rri8 : Ii8<0xAC, MRMDestReg,
+                     (ops R16:$dst, R16:$src1, R16:$src2, i8imm:$src3),
+                     "shrd{w} {$src3, $src2, $dst|$dst, $src2, $src3}">,
+                     TB, OpSize;
 }
 
 let isTwoAddress = 0 in {
@@ -904,6 +918,21 @@ let isTwoAddress = 0 in {
   def SHRD32mri8 : Ii8<0xAC, MRMDestMem, 
                        (ops i32mem:$dst, R32:$src2, i8imm:$src3),
                        "shrd{l} {$src3, $src2, $dst|$dst, $src2, $src3}">, TB;
+
+  def SHLD16mrCL : I<0xA5, MRMDestMem, (ops i16mem:$dst, R16:$src2),
+                     "shld{w} {%cl, $src2, $dst|$dst, $src2, %CL}">,
+                     Imp<[CL],[]>, TB, OpSize;
+  def SHRD16mrCL : I<0xAD, MRMDestMem, (ops i16mem:$dst, R16:$src2),
+                    "shrd{w} {%cl, $src2, $dst|$dst, $src2, %CL}">,
+                    Imp<[CL],[]>, TB, OpSize;
+  def SHLD16mri8 : Ii8<0xA4, MRMDestMem,
+                      (ops i16mem:$dst, R16:$src2, i8imm:$src3),
+                      "shld{w} {$src3, $src2, $dst|$dst, $src2, $src3}">,
+                      TB, OpSize;
+  def SHRD16mri8 : Ii8<0xAC, MRMDestMem, 
+                       (ops i16mem:$dst, R16:$src2, i8imm:$src3),
+                       "shrd{w} {$src3, $src2, $dst|$dst, $src2, $src3}">,
+                       TB, OpSize;
 }
 
 
diff --git a/lib/Target/X86/X86RegisterInfo.cpp b/lib/Target/X86/X86RegisterInfo.cpp
index 456c0b32d09..88e83380588 100644
--- a/lib/Target/X86/X86RegisterInfo.cpp
+++ b/lib/Target/X86/X86RegisterInfo.cpp
@@ -225,6 +225,10 @@ MachineInstr* X86RegisterInfo::foldMemoryOperand(MachineInstr* MI,
     case X86::SHLD32rri8:return MakeMRIInst(X86::SHLD32mri8,FrameIndex, MI);
     case X86::SHRD32rrCL:return MakeMRInst( X86::SHRD32mrCL,FrameIndex, MI);
     case X86::SHRD32rri8:return MakeMRIInst(X86::SHRD32mri8,FrameIndex, MI);
+    case X86::SHLD16rrCL:return MakeMRInst( X86::SHLD16mrCL,FrameIndex, MI);
+    case X86::SHLD16rri8:return MakeMRIInst(X86::SHLD16mri8,FrameIndex, MI);
+    case X86::SHRD16rrCL:return MakeMRInst( X86::SHRD16mrCL,FrameIndex, MI);
+    case X86::SHRD16rri8:return MakeMRIInst(X86::SHRD16mri8,FrameIndex, MI);
     case X86::SETBr:     return MakeMInst( X86::SETBm,   FrameIndex, MI);
     case X86::SETAEr:    return MakeMInst( X86::SETAEm,  FrameIndex, MI);
     case X86::SETEr:     return MakeMInst( X86::SETEm,   FrameIndex, MI);