diff --git a/lib/Target/ARM/ARMInstrFormats.td b/lib/Target/ARM/ARMInstrFormats.td
index 3d43edba728..7069e91ce54 100644
--- a/lib/Target/ARM/ARMInstrFormats.td
+++ b/lib/Target/ARM/ARMInstrFormats.td
@@ -1754,6 +1754,17 @@ class N2VImm<bit op24, bit op23, bits<4> op11_8, bit op7, bit op6, bit op4,
   let Inst{7}    = op7;
   let Inst{6}    = op6;
   let Inst{4}    = op4;
+  
+  // Instruction operands.
+  bits<5> Vd;
+  bits<5> Vm;
+  bits<6> SIMM;
+
+  let Inst{15-12} = Vd{3-0};
+  let Inst{22}    = Vd{4};
+  let Inst{3-0}   = Vm{3-0};
+  let Inst{5}     = Vm{4};
+  let Inst{21-16} = SIMM{5-0};
 }
 
 // NEON 3 vector register format.
diff --git a/lib/Target/ARM/ARMInstrNEON.td b/lib/Target/ARM/ARMInstrNEON.td
index 232dd486a05..afb2afa7d60 100644
--- a/lib/Target/ARM/ARMInstrNEON.td
+++ b/lib/Target/ARM/ARMInstrNEON.td
@@ -1289,6 +1289,15 @@ class N3VDIntSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin,
                          (Ty (NEONvduplane (Ty DPR_8:$src2), imm:$lane)))))]> {
   let isCommutable = 0;
 }
+class N3VDIntSh<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4,
+              Format f, InstrItinClass itin, string OpcodeStr, string Dt,
+              ValueType ResTy, ValueType OpTy, Intrinsic IntOp, bit Commutable>
+  : N3V<op24, op23, op21_20, op11_8, 0, op4,
+        (outs DPR:$Vd), (ins DPR:$Vm, DPR:$Vn), f, itin,
+        OpcodeStr, Dt, "$Vd, $Vm, $Vn", "",
+        [(set DPR:$Vd, (ResTy (IntOp (OpTy DPR:$Vm), (OpTy DPR:$Vn))))]> {
+  let isCommutable = Commutable;
+}
 
 class N3VQInt<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4,
               Format f, InstrItinClass itin, string OpcodeStr, string Dt,
@@ -1323,6 +1332,15 @@ class N3VQIntSL16<bits<2> op21_20, bits<4> op11_8, InstrItinClass itin,
                                                  imm:$lane)))))]> {
   let isCommutable = 0;
 }
+class N3VQIntSh<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4,
+              Format f, InstrItinClass itin, string OpcodeStr, string Dt,
+              ValueType ResTy, ValueType OpTy, Intrinsic IntOp, bit Commutable>
+  : N3V<op24, op23, op21_20, op11_8, 1, op4,
+        (outs QPR:$Vd), (ins QPR:$Vm, QPR:$Vn), f, itin,
+        OpcodeStr, Dt, "$Vd, $Vm, $Vn", "",
+        [(set QPR:$Vd, (ResTy (IntOp (OpTy QPR:$Vm), (OpTy QPR:$Vn))))]> {
+  let isCommutable = Commutable;
+}
 
 // Multiply-Add/Sub operations: single-, double- and quad-register.
 class N3VSMulOp<bit op24, bit op23, bits<2> op21_20, bits<4> op11_8, bit op4,
@@ -1936,6 +1954,27 @@ multiclass N3VInt_HS<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
                       OpcodeStr, !strconcat(Dt, "32"),
                       v4i32, v4i32, IntOp, Commutable>;
 }
+multiclass N3VInt_HSSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
+                     InstrItinClass itinD16, InstrItinClass itinD32,
+                     InstrItinClass itinQ16, InstrItinClass itinQ32,
+                     string OpcodeStr, string Dt,
+                     Intrinsic IntOp, bit Commutable = 0> {
+  // 64-bit vector types.
+  def v4i16 : N3VDIntSh<op24, op23, 0b01, op11_8, op4, f, itinD16,
+                      OpcodeStr, !strconcat(Dt, "16"),
+                      v4i16, v4i16, IntOp, Commutable>;
+  def v2i32 : N3VDIntSh<op24, op23, 0b10, op11_8, op4, f, itinD32,
+                      OpcodeStr, !strconcat(Dt, "32"),
+                      v2i32, v2i32, IntOp, Commutable>;
+
+  // 128-bit vector types.
+  def v8i16 : N3VQIntSh<op24, op23, 0b01, op11_8, op4, f, itinQ16,
+                      OpcodeStr, !strconcat(Dt, "16"),
+                      v8i16, v8i16, IntOp, Commutable>;
+  def v4i32 : N3VQIntSh<op24, op23, 0b10, op11_8, op4, f, itinQ32,
+                      OpcodeStr, !strconcat(Dt, "32"),
+                      v4i32, v4i32, IntOp, Commutable>;
+}
 
 multiclass N3VIntSL_HS<bits<4> op11_8, 
                        InstrItinClass itinD16, InstrItinClass itinD32,
@@ -1966,6 +2005,21 @@ multiclass N3VInt_QHS<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
                       OpcodeStr, !strconcat(Dt, "8"),
                       v16i8, v16i8, IntOp, Commutable>;
 }
+multiclass N3VInt_QHSSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
+                      InstrItinClass itinD16, InstrItinClass itinD32,
+                      InstrItinClass itinQ16, InstrItinClass itinQ32,
+                      string OpcodeStr, string Dt,
+                      Intrinsic IntOp, bit Commutable = 0>
+  : N3VInt_HSSh<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32,
+              OpcodeStr, Dt, IntOp, Commutable> {
+  def v8i8  : N3VDIntSh<op24, op23, 0b00, op11_8, op4, f, itinD16,
+                      OpcodeStr, !strconcat(Dt, "8"),
+                      v8i8, v8i8, IntOp, Commutable>;
+  def v16i8 : N3VQIntSh<op24, op23, 0b00, op11_8, op4, f, itinQ16,
+                      OpcodeStr, !strconcat(Dt, "8"),
+                      v16i8, v16i8, IntOp, Commutable>;
+}
+
 
 // ....then also with element size of 64 bits:
 multiclass N3VInt_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
@@ -1982,6 +2036,20 @@ multiclass N3VInt_QHSD<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
                       OpcodeStr, !strconcat(Dt, "64"),
                       v2i64, v2i64, IntOp, Commutable>;
 }
+multiclass N3VInt_QHSDSh<bit op24, bit op23, bits<4> op11_8, bit op4, Format f,
+                       InstrItinClass itinD16, InstrItinClass itinD32,
+                       InstrItinClass itinQ16, InstrItinClass itinQ32,
+                       string OpcodeStr, string Dt,
+                       Intrinsic IntOp, bit Commutable = 0>
+  : N3VInt_QHSSh<op24, op23, op11_8, op4, f, itinD16, itinD32, itinQ16, itinQ32,
+               OpcodeStr, Dt, IntOp, Commutable> {
+  def v1i64 : N3VDIntSh<op24, op23, 0b11, op11_8, op4, f, itinD32,
+                      OpcodeStr, !strconcat(Dt, "64"),
+                      v1i64, v1i64, IntOp, Commutable>;
+  def v2i64 : N3VQIntSh<op24, op23, 0b11, op11_8, op4, f, itinQ32,
+                      OpcodeStr, !strconcat(Dt, "64"),
+                      v2i64, v2i64, IntOp, Commutable>;
+}
 
 // Neon Narrowing 3-register vector intrinsics,
 //   source operand element sizes of 16, 32 and 64 bits:
@@ -3160,10 +3228,10 @@ def VRSQRTSfq : N3VQInt<0, 0, 0b10, 0b1111, 1, N3RegFrm,
 // Vector Shifts.
 
 //   VSHL     : Vector Shift
-defm VSHLs    : N3VInt_QHSD<0, 0, 0b0100, 0, N3RegVShFrm,
+defm VSHLs    : N3VInt_QHSDSh<0, 0, 0b0100, 0, N3RegVShFrm,
                             IIC_VSHLiD, IIC_VSHLiD, IIC_VSHLiQ, IIC_VSHLiQ,
                             "vshl", "s", int_arm_neon_vshifts, 0>;
-defm VSHLu    : N3VInt_QHSD<1, 0, 0b0100, 0, N3RegVShFrm,
+defm VSHLu    : N3VInt_QHSDSh<1, 0, 0b0100, 0, N3RegVShFrm,
                             IIC_VSHLiD, IIC_VSHLiD, IIC_VSHLiQ, IIC_VSHLiQ,
                             "vshl", "u", int_arm_neon_vshiftu, 0>;
 //   VSHL     : Vector Shift Left (Immediate)
diff --git a/test/MC/ARM/neon-shift-encoding.ll b/test/MC/ARM/neon-shift-encoding.ll
new file mode 100644
index 00000000000..0c45b1480c1
--- /dev/null
+++ b/test/MC/ARM/neon-shift-encoding.ll
@@ -0,0 +1,136 @@
+; RUN: llc -show-mc-encoding -march=arm -mcpu=cortex-a8 -mattr=+neon < %s | FileCheck %s
+
+; CHECK: vshls_8xi8
+define <8 x i8> @vshls_8xi8(<8 x i8>* %A, <8 x i8>* %B) nounwind {
+	%tmp1 = load <8 x i8>* %A
+	%tmp2 = load <8 x i8>* %B
+; CHECK: vshl.u8	d16, d17, d16           @ encoding: [0xa1,0x04,0x40,0xf3]
+	%tmp3 = shl <8 x i8> %tmp1, %tmp2
+	ret <8 x i8> %tmp3
+}
+
+; CHECK: vshls_4xi16
+define <4 x i16> @vshls_4xi16(<4 x i16>* %A, <4 x i16>* %B) nounwind {
+	%tmp1 = load <4 x i16>* %A
+	%tmp2 = load <4 x i16>* %B
+; CHECK: vshl.u16	d16, d17, d16   @ encoding: [0xa1,0x04,0x50,0xf3]
+	%tmp3 = shl <4 x i16> %tmp1, %tmp2
+	ret <4 x i16> %tmp3
+}
+
+; CHECK: vshls_2xi32
+define <2 x i32> @vshls_2xi32(<2 x i32>* %A, <2 x i32>* %B) nounwind {
+	%tmp1 = load <2 x i32>* %A
+	%tmp2 = load <2 x i32>* %B
+; CHECK: vshl.u32	d16, d17, d16   @ encoding: [0xa1,0x04,0x60,0xf3]
+	%tmp3 = shl <2 x i32> %tmp1, %tmp2
+	ret <2 x i32> %tmp3
+}
+
+; CHECK: vshls_1xi64
+define <1 x i64> @vshls_1xi64(<1 x i64>* %A, <1 x i64>* %B) nounwind {
+	%tmp1 = load <1 x i64>* %A
+	%tmp2 = load <1 x i64>* %B
+; CHECK: vshl.u64	d16, d17, d16   @ encoding: [0xa1,0x04,0x70,0xf3]
+	%tmp3 = shl <1 x i64> %tmp1, %tmp2
+	ret <1 x i64> %tmp3
+}
+
+; CHECK: vshli_8xi8
+define <8 x i8> @vshli_8xi8(<8 x i8>* %A) nounwind {
+	%tmp1 = load <8 x i8>* %A
+; CHECK: vshl.i8	d16, d16, #7            @ encoding: [0x30,0x05,0xcf,0xf2]
+	%tmp2 = shl <8 x i8> %tmp1, < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >
+	ret <8 x i8> %tmp2
+}
+
+; CHECK: vshli_4xi16
+define <4 x i16> @vshli_4xi16(<4 x i16>* %A) nounwind {
+	%tmp1 = load <4 x i16>* %A
+; CHECK: vshl.i16	d16, d16, #15   @ encoding: [0x30,0x05,0xdf,0xf2
+	%tmp2 = shl <4 x i16> %tmp1, < i16 15, i16 15, i16 15, i16 15 >
+	ret <4 x i16> %tmp2
+}
+
+; CHECK: vshli_2xi32
+define <2 x i32> @vshli_2xi32(<2 x i32>* %A) nounwind {
+	%tmp1 = load <2 x i32>* %A
+; CHECK: vshl.i32	d16, d16, #31   @ encoding: [0x30,0x05,0xff,0xf2]
+	%tmp2 = shl <2 x i32> %tmp1, < i32 31, i32 31 >
+	ret <2 x i32> %tmp2
+}
+
+; CHECK: vshli_1xi64
+define <1 x i64> @vshli_1xi64(<1 x i64>* %A) nounwind {
+	%tmp1 = load <1 x i64>* %A
+; CHECK: vshl.i64	d16, d16, #63   @ encoding: [0xb0,0x05,0xff,0xf2]
+	%tmp2 = shl <1 x i64> %tmp1, < i64 63 >
+	ret <1 x i64> %tmp2
+}
+
+; CHECK: vshls_16xi8
+define <16 x i8> @vshls_16xi8(<16 x i8>* %A, <16 x i8>* %B) nounwind {
+	%tmp1 = load <16 x i8>* %A
+	%tmp2 = load <16 x i8>* %B
+; CHECK: vshl.u8	q8, q8, q9              @ encoding: [0xe0,0x04,0x42,0xf3]
+	%tmp3 = shl <16 x i8> %tmp1, %tmp2
+	ret <16 x i8> %tmp3
+}
+
+; CHECK: vshls_8xi16
+define <8 x i16> @vshls_8xi16(<8 x i16>* %A, <8 x i16>* %B) nounwind {
+	%tmp1 = load <8 x i16>* %A
+	%tmp2 = load <8 x i16>* %B
+	%tmp3 = shl <8 x i16> %tmp1, %tmp2
+	ret <8 x i16> %tmp3
+}
+
+; CHECK: vshls_4xi32
+define <4 x i32> @vshls_4xi32(<4 x i32>* %A, <4 x i32>* %B) nounwind {
+	%tmp1 = load <4 x i32>* %A
+	%tmp2 = load <4 x i32>* %B
+; CHECK: vshl.u32	q8, q8, q9      @ encoding: [0xe0,0x04,0x62,0xf3]
+	%tmp3 = shl <4 x i32> %tmp1, %tmp2
+	ret <4 x i32> %tmp3
+}
+
+; CHECK: vshls_2xi64
+define <2 x i64> @vshls_2xi64(<2 x i64>* %A, <2 x i64>* %B) nounwind {
+	%tmp1 = load <2 x i64>* %A
+	%tmp2 = load <2 x i64>* %B
+; CHECK: vshl.u64	q8, q8, q9      @ encoding: [0xe0,0x04,0x72,0xf3]
+	%tmp3 = shl <2 x i64> %tmp1, %tmp2
+	ret <2 x i64> %tmp3
+}
+
+; CHECK: vshli_16xi8
+define <16 x i8> @vshli_16xi8(<16 x i8>* %A) nounwind {
+	%tmp1 = load <16 x i8>* %A
+; CHECK: vshl.i8	q8, q8, #7              @ encoding: [0x70,0x05,0xcf,0xf2]
+	%tmp2 = shl <16 x i8> %tmp1, < i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7, i8 7 >
+	ret <16 x i8> %tmp2
+}
+
+; CHECK: vshli_8xi16
+define <8 x i16> @vshli_8xi16(<8 x i16>* %A) nounwind {
+	%tmp1 = load <8 x i16>* %A
+; CHECK: vshl.i16	q8, q8, #15     @ encoding: [0x70,0x05,0xdf,0xf2]
+	%tmp2 = shl <8 x i16> %tmp1, < i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15 >
+	ret <8 x i16> %tmp2
+}
+
+; CHECK: vshli_4xi32
+define <4 x i32> @vshli_4xi32(<4 x i32>* %A) nounwind {
+	%tmp1 = load <4 x i32>* %A
+; CHECK: vshl.i32	q8, q8, #31     @ encoding: [0x70,0x05,0xff,0xf2]
+	%tmp2 = shl <4 x i32> %tmp1, < i32 31, i32 31, i32 31, i32 31 >
+	ret <4 x i32> %tmp2
+}
+
+; CHECK: vshli_2xi64
+define <2 x i64> @vshli_2xi64(<2 x i64>* %A) nounwind {
+	%tmp1 = load <2 x i64>* %A
+; CHECK: vshl.i64	q8, q8, #63     @ encoding: [0xf0,0x05,0xff,0xf2]
+	%tmp2 = shl <2 x i64> %tmp1, < i64 63, i64 63 >
+	ret <2 x i64> %tmp2
+}