diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -39593,6 +39593,7 @@
   SDValue N0 = N->getOperand(0);
   EVT VT = N->getValueType(0);
   EVT SrcVT = N0.getValueType();
+  const TargetLowering &TLI = DAG.getTargetLoweringInfo();
 
   // Try to match patterns such as
   // (i16 bitcast (v16i1 x))
@@ -39650,8 +39651,7 @@
     // If we're bitcasting from iX to vXi1, see if the integer originally
     // began as a vXi1 and whether we can remove the bitcast entirely.
     if (VT.isVector() && VT.getScalarType() == MVT::i1 &&
-        SrcVT.isScalarInteger() &&
-        DAG.getTargetLoweringInfo().isTypeLegal(VT)) {
+        SrcVT.isScalarInteger() && TLI.isTypeLegal(VT)) {
       if (SDValue V =
               combineBitcastToBoolVector(VT, N0, SDLoc(N), DAG, Subtarget))
         return V;
@@ -39839,8 +39839,11 @@
     default: return SDValue();
   }
 
+  // Check if we have a bitcast from another integer type as well.
   if (!((Subtarget.hasSSE1() && VT == MVT::f32) ||
-        (Subtarget.hasSSE2() && VT == MVT::f64)))
+        (Subtarget.hasSSE2() && VT == MVT::f64) ||
+        (Subtarget.hasSSE2() && VT.isInteger() && VT.isVector() &&
+         TLI.isTypeLegal(VT))))
     return SDValue();
 
   SDValue LogicOp0 = N0.getOperand(0);
@@ -39849,17 +39852,21 @@
 
   // bitcast(logic(bitcast(X), Y)) --> logic'(X, bitcast(Y))
   if (N0.hasOneUse() && LogicOp0.getOpcode() == ISD::BITCAST &&
-      LogicOp0.hasOneUse() && LogicOp0.getOperand(0).getValueType() == VT &&
+      LogicOp0.hasOneUse() && LogicOp0.getOperand(0).hasOneUse() &&
+      LogicOp0.getOperand(0).getValueType() == VT &&
       !isa<ConstantSDNode>(LogicOp0.getOperand(0))) {
     SDValue CastedOp1 = DAG.getBitcast(VT, LogicOp1);
-    return DAG.getNode(FPOpcode, DL0, VT, LogicOp0.getOperand(0), CastedOp1);
+    unsigned Opcode = VT.isFloatingPoint() ? FPOpcode : N0.getOpcode();
+    return DAG.getNode(Opcode, DL0, VT, LogicOp0.getOperand(0), CastedOp1);
   }
   // bitcast(logic(X, bitcast(Y))) --> logic'(bitcast(X), Y)
   if (N0.hasOneUse() && LogicOp1.getOpcode() == ISD::BITCAST &&
-      LogicOp1.hasOneUse() && LogicOp1.getOperand(0).getValueType() == VT &&
+      LogicOp1.hasOneUse() && LogicOp1.getOperand(0).hasOneUse() &&
+      LogicOp1.getOperand(0).getValueType() == VT &&
       !isa<ConstantSDNode>(LogicOp1.getOperand(0))) {
     SDValue CastedOp0 = DAG.getBitcast(VT, LogicOp0);
-    return DAG.getNode(FPOpcode, DL0, VT, LogicOp1.getOperand(0), CastedOp0);
+    unsigned Opcode = VT.isFloatingPoint() ? FPOpcode : N0.getOpcode();
+    return DAG.getNode(Opcode, DL0, VT, LogicOp1.getOperand(0), CastedOp0);
   }
 
   return SDValue();
@@ -49335,6 +49342,29 @@
     }
   }
 
+  // concat(extract_subvector(v0,c0), extract_subvector(v1,c1)) -> vperm2x128.
+  // Only concat of subvector high halves which vperm2x128 is best at.
+  // TODO: This should go in combineX86ShufflesRecursively eventually.
+  if (VT.is256BitVector() && Ops.size() == 2) {
+    SDValue Src0 = peekThroughBitcasts(Ops[0]);
+    SDValue Src1 = peekThroughBitcasts(Ops[1]);
+    if (Src0.getOpcode() == ISD::EXTRACT_SUBVECTOR &&
+        Src1.getOpcode() == ISD::EXTRACT_SUBVECTOR) {
+      EVT SrcVT0 = Src0.getOperand(0).getValueType();
+      EVT SrcVT1 = Src1.getOperand(0).getValueType();
+      unsigned NumSrcElts0 = SrcVT0.getVectorNumElements();
+      unsigned NumSrcElts1 = SrcVT1.getVectorNumElements();
+      if (SrcVT0.is256BitVector() && SrcVT1.is256BitVector() &&
+          Src0.getConstantOperandAPInt(1) == (NumSrcElts0 / 2) &&
+          Src1.getConstantOperandAPInt(1) == (NumSrcElts1 / 2)) {
+        return DAG.getNode(X86ISD::VPERM2X128, DL, VT,
+                           DAG.getBitcast(VT, Src0.getOperand(0)),
+                           DAG.getBitcast(VT, Src1.getOperand(0)),
+                           DAG.getTargetConstant(0x31, DL, MVT::i8));
+      }
+    }
+  }
+
   // Repeated opcode.
   // TODO - combineX86ShufflesRecursively should handle shuffle concatenation
   // but it currently struggles with different vector widths.
diff --git a/llvm/test/CodeGen/X86/avx512-logic.ll b/llvm/test/CodeGen/X86/avx512-logic.ll
--- a/llvm/test/CodeGen/X86/avx512-logic.ll
+++ b/llvm/test/CodeGen/X86/avx512-logic.ll
@@ -911,7 +911,7 @@
 ; KNL:       ## %bb.0:
 ; KNL-NEXT:    vpxor %xmm3, %xmm3, %xmm3
 ; KNL-NEXT:    vpcmpgtd %zmm2, %zmm3, %k1
-; KNL-NEXT:    vpandq {{.*}}(%rip), %zmm0, %zmm0
+; KNL-NEXT:    vpandd {{.*}}(%rip), %zmm0, %zmm0
 ; KNL-NEXT:    vpord %zmm1, %zmm0, %zmm0 {%k1} {z}
 ; KNL-NEXT:    retq
 ;
@@ -933,7 +933,7 @@
 ; KNL:       ## %bb.0:
 ; KNL-NEXT:    vpxor %xmm3, %xmm3, %xmm3
 ; KNL-NEXT:    vpcmpgtq %zmm2, %zmm3, %k1
-; KNL-NEXT:    vpandd {{.*}}(%rip), %zmm0, %zmm0
+; KNL-NEXT:    vpandq {{.*}}(%rip), %zmm0, %zmm0
 ; KNL-NEXT:    vpxorq %zmm1, %zmm0, %zmm0 {%k1} {z}
 ; KNL-NEXT:    retq
 ;
@@ -977,7 +977,7 @@
 ; KNL:       ## %bb.0:
 ; KNL-NEXT:    vpxor %xmm3, %xmm3, %xmm3
 ; KNL-NEXT:    vpcmpgtd %zmm2, %zmm3, %k1
-; KNL-NEXT:    vpandq {{.*}}(%rip), %zmm0, %zmm0
+; KNL-NEXT:    vpandd {{.*}}(%rip), %zmm0, %zmm0
 ; KNL-NEXT:    vpord %zmm1, %zmm0, %zmm1 {%k1}
 ; KNL-NEXT:    vmovdqa64 %zmm1, %zmm0
 ; KNL-NEXT:    retq
@@ -1023,7 +1023,7 @@
 ; KNL:       ## %bb.0:
 ; KNL-NEXT:    vpxor %xmm3, %xmm3, %xmm3
 ; KNL-NEXT:    vpcmpgtq %zmm2, %zmm3, %k1
-; KNL-NEXT:    vpandd {{.*}}(%rip), %zmm0, %zmm0
+; KNL-NEXT:    vpandq {{.*}}(%rip), %zmm0, %zmm0
 ; KNL-NEXT:    vpxorq %zmm1, %zmm0, %zmm1 {%k1}
 ; KNL-NEXT:    vmovdqa64 %zmm1, %zmm0
 ; KNL-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/known-signbits-vector.ll b/llvm/test/CodeGen/X86/known-signbits-vector.ll
--- a/llvm/test/CodeGen/X86/known-signbits-vector.ll
+++ b/llvm/test/CodeGen/X86/known-signbits-vector.ll
@@ -341,7 +341,7 @@
 ; X86-NEXT:    vpsrad $29, %xmm0, %xmm0
 ; X86-NEXT:    vpshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; X86-NEXT:    vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X86-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; X86-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; X86-NEXT:    vcvtdq2ps %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
diff --git a/llvm/test/CodeGen/X86/memcmp-more-load-pairs.ll b/llvm/test/CodeGen/X86/memcmp-more-load-pairs.ll
--- a/llvm/test/CodeGen/X86/memcmp-more-load-pairs.ll
+++ b/llvm/test/CodeGen/X86/memcmp-more-load-pairs.ll
@@ -1671,7 +1671,7 @@
 ; X64-AVX-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; X64-AVX-NEXT:    vpxor %xmm2, %xmm1, %xmm1
 ; X64-AVX-NEXT:    vpxor (%rsi), %xmm0, %xmm0
-; X64-AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; X64-AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; X64-AVX-NEXT:    vptest %xmm0, %xmm0
 ; X64-AVX-NEXT:    sete %al
 ; X64-AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/memcmp-optsize.ll b/llvm/test/CodeGen/X86/memcmp-optsize.ll
--- a/llvm/test/CodeGen/X86/memcmp-optsize.ll
+++ b/llvm/test/CodeGen/X86/memcmp-optsize.ll
@@ -688,7 +688,7 @@
 ; X64-AVX-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; X64-AVX-NEXT:    vpxor %xmm2, %xmm1, %xmm1
 ; X64-AVX-NEXT:    vpxor (%rsi), %xmm0, %xmm0
-; X64-AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; X64-AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; X64-AVX-NEXT:    vptest %xmm0, %xmm0
 ; X64-AVX-NEXT:    sete %al
 ; X64-AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/memcmp-pgso.ll b/llvm/test/CodeGen/X86/memcmp-pgso.ll
--- a/llvm/test/CodeGen/X86/memcmp-pgso.ll
+++ b/llvm/test/CodeGen/X86/memcmp-pgso.ll
@@ -688,7 +688,7 @@
 ; X64-AVX-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; X64-AVX-NEXT:    vpxor %xmm2, %xmm1, %xmm1
 ; X64-AVX-NEXT:    vpxor (%rsi), %xmm0, %xmm0
-; X64-AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; X64-AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; X64-AVX-NEXT:    vptest %xmm0, %xmm0
 ; X64-AVX-NEXT:    sete %al
 ; X64-AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/memcmp.ll b/llvm/test/CodeGen/X86/memcmp.ll
--- a/llvm/test/CodeGen/X86/memcmp.ll
+++ b/llvm/test/CodeGen/X86/memcmp.ll
@@ -1650,7 +1650,7 @@
 ; X64-AVX-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
 ; X64-AVX-NEXT:    vpxor %xmm2, %xmm1, %xmm1
 ; X64-AVX-NEXT:    vpxor (%rsi), %xmm0, %xmm0
-; X64-AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; X64-AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; X64-AVX-NEXT:    vptest %xmm0, %xmm0
 ; X64-AVX-NEXT:    sete %al
 ; X64-AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll b/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
--- a/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
+++ b/llvm/test/CodeGen/X86/merge-consecutive-loads-512.ll
@@ -217,7 +217,7 @@
 ; X86-AVX512F:       # %bb.0:
 ; X86-AVX512F-NEXT:    movl {{[0-9]+}}(%esp), %eax
 ; X86-AVX512F-NEXT:    vmovdqu64 8(%eax), %zmm0
-; X86-AVX512F-NEXT:    vpandq {{\.LCPI.*}}, %zmm0, %zmm0
+; X86-AVX512F-NEXT:    vpandd {{\.LCPI.*}}, %zmm0, %zmm0
 ; X86-AVX512F-NEXT:    retl
   %ptr0 = getelementptr inbounds i64, i64* %ptr, i64 1
   %ptr2 = getelementptr inbounds i64, i64* %ptr, i64 3
diff --git a/llvm/test/CodeGen/X86/sse2.ll b/llvm/test/CodeGen/X86/sse2.ll
--- a/llvm/test/CodeGen/X86/sse2.ll
+++ b/llvm/test/CodeGen/X86/sse2.ll
@@ -678,38 +678,16 @@
 ; X86-SSE-NEXT:    andps {{\.LCPI.*}}, %xmm0
 ; X86-SSE-NEXT:    retl
 ;
-; X86-AVX-LABEL: PR19721:
-; X86-AVX:       # %bb.0:
-; X86-AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
-; X86-AVX-NEXT:    vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
-; X86-AVX-NEXT:    retl
+; AVX-LABEL: PR19721:
+; AVX:       # %bb.0:
+; AVX-NEXT:    vxorps %xmm1, %xmm1, %xmm1
+; AVX-NEXT:    vblendps {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
+; AVX-NEXT:    ret{{[l|q]}}
 ;
 ; X64-SSE-LABEL: PR19721:
 ; X64-SSE:       # %bb.0:
-; X64-SSE-NEXT:    movq %xmm0, %rax
-; X64-SSE-NEXT:    movabsq $-4294967296, %rcx # imm = 0xFFFFFFFF00000000
-; X64-SSE-NEXT:    andq %rax, %rcx
-; X64-SSE-NEXT:    movq %rcx, %xmm1
-; X64-SSE-NEXT:    movsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X64-SSE-NEXT:    andps {{.*}}(%rip), %xmm0
 ; X64-SSE-NEXT:    retq
-;
-; X64-AVX1-LABEL: PR19721:
-; X64-AVX1:       # %bb.0:
-; X64-AVX1-NEXT:    vmovq %xmm0, %rax
-; X64-AVX1-NEXT:    movabsq $-4294967296, %rcx # imm = 0xFFFFFFFF00000000
-; X64-AVX1-NEXT:    andq %rax, %rcx
-; X64-AVX1-NEXT:    vmovq %rcx, %xmm1
-; X64-AVX1-NEXT:    vpblendw {{.*#+}} xmm0 = xmm1[0,1,2,3],xmm0[4,5,6,7]
-; X64-AVX1-NEXT:    retq
-;
-; X64-AVX512-LABEL: PR19721:
-; X64-AVX512:       # %bb.0:
-; X64-AVX512-NEXT:    vmovq %xmm0, %rax
-; X64-AVX512-NEXT:    movabsq $-4294967296, %rcx # imm = 0xFFFFFFFF00000000
-; X64-AVX512-NEXT:    andq %rax, %rcx
-; X64-AVX512-NEXT:    vmovq %rcx, %xmm1
-; X64-AVX512-NEXT:    vpblendd {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3]
-; X64-AVX512-NEXT:    retq
   %bc = bitcast <4 x i32> %i to i128
   %insert = and i128 %bc, -4294967296
   %bc2 = bitcast i128 %insert to <4 x i32>
diff --git a/llvm/test/CodeGen/X86/vec_saddo.ll b/llvm/test/CodeGen/X86/vec_saddo.ll
--- a/llvm/test/CodeGen/X86/vec_saddo.ll
+++ b/llvm/test/CodeGen/X86/vec_saddo.ll
@@ -813,10 +813,9 @@
 ; SSE-NEXT:    pxor %xmm0, %xmm2
 ; SSE-NEXT:    movdqa %xmm3, %xmm4
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm4
-; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
 ; SSE-NEXT:    pcmpeqd %xmm3, %xmm2
 ; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
-; SSE-NEXT:    pand %xmm5, %xmm2
+; SSE-NEXT:    pand %xmm4, %xmm2
 ; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
 ; SSE-NEXT:    por %xmm2, %xmm3
 ; SSE-NEXT:    pxor %xmm2, %xmm2
diff --git a/llvm/test/CodeGen/X86/vec_ssubo.ll b/llvm/test/CodeGen/X86/vec_ssubo.ll
--- a/llvm/test/CodeGen/X86/vec_ssubo.ll
+++ b/llvm/test/CodeGen/X86/vec_ssubo.ll
@@ -819,19 +819,17 @@
 ; SSE-NEXT:    pxor %xmm2, %xmm0
 ; SSE-NEXT:    movdqa %xmm3, %xmm4
 ; SSE-NEXT:    pcmpgtd %xmm0, %xmm4
-; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
 ; SSE-NEXT:    pcmpeqd %xmm3, %xmm0
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE-NEXT:    pand %xmm5, %xmm0
+; SSE-NEXT:    pand %xmm4, %xmm0
 ; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[1,1,3,3]
 ; SSE-NEXT:    por %xmm0, %xmm3
 ; SSE-NEXT:    pxor %xmm2, %xmm1
 ; SSE-NEXT:    movdqa %xmm1, %xmm0
 ; SSE-NEXT:    pcmpgtd %xmm2, %xmm0
-; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,0,2,2]
 ; SSE-NEXT:    pcmpeqd %xmm2, %xmm1
 ; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; SSE-NEXT:    pand %xmm4, %xmm1
+; SSE-NEXT:    pand %xmm0, %xmm1
 ; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
 ; SSE-NEXT:    por %xmm1, %xmm0
 ; SSE-NEXT:    pxor %xmm3, %xmm0
diff --git a/llvm/test/CodeGen/X86/vector-idiv-v2i32.ll b/llvm/test/CodeGen/X86/vector-idiv-v2i32.ll
--- a/llvm/test/CodeGen/X86/vector-idiv-v2i32.ll
+++ b/llvm/test/CodeGen/X86/vector-idiv-v2i32.ll
@@ -249,9 +249,9 @@
 define void @test_urem_pow2_v2i32(<2 x i32>* %x, <2 x i32>* %y) nounwind {
 ; X64-LABEL: test_urem_pow2_v2i32:
 ; X64:       # %bb.0:
-; X64-NEXT:    movsd {{.*#+}} xmm0 = mem[0],zero
-; X64-NEXT:    andps {{.*}}(%rip), %xmm0
-; X64-NEXT:    movlps %xmm0, (%rsi)
+; X64-NEXT:    movabsq $30064771079, %rax # imm = 0x700000007
+; X64-NEXT:    andq (%rdi), %rax
+; X64-NEXT:    movq %rax, (%rsi)
 ; X64-NEXT:    retq
 ;
 ; X86-LABEL: test_urem_pow2_v2i32:
diff --git a/llvm/test/CodeGen/X86/vector-reduce-and-cmp.ll b/llvm/test/CodeGen/X86/vector-reduce-and-cmp.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-and-cmp.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-and-cmp.ll
@@ -477,7 +477,7 @@
 ; AVX-LABEL: test_v2i16:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    testw %ax, %ax
 ; AVX-NEXT:    sete %al
@@ -505,7 +505,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    testw %ax, %ax
 ; AVX-NEXT:    setne %al
@@ -537,7 +537,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    testw %ax, %ax
 ; AVX-NEXT:    sete %al
diff --git a/llvm/test/CodeGen/X86/vector-reduce-and.ll b/llvm/test/CodeGen/X86/vector-reduce-and.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-and.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-and.ll
@@ -407,7 +407,7 @@
 ; AVX-LABEL: test_v2i16:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -432,7 +432,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -461,7 +461,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpand %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpand %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-reduce-or.ll b/llvm/test/CodeGen/X86/vector-reduce-or.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-or.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-or.ll
@@ -407,7 +407,7 @@
 ; AVX-LABEL: test_v2i16:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -432,7 +432,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -461,7 +461,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-reduce-xor.ll b/llvm/test/CodeGen/X86/vector-reduce-xor.ll
--- a/llvm/test/CodeGen/X86/vector-reduce-xor.ll
+++ b/llvm/test/CodeGen/X86/vector-reduce-xor.ll
@@ -407,7 +407,7 @@
 ; AVX-LABEL: test_v2i16:
 ; AVX:       # %bb.0:
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpxor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpxor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -432,7 +432,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpxor %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpxor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpxor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
@@ -461,7 +461,7 @@
 ; AVX-NEXT:    vpshufd {{.*#+}} xmm1 = xmm0[1,1,1,1]
 ; AVX-NEXT:    vpxor %xmm1, %xmm0, %xmm0
 ; AVX-NEXT:    vpsrld $16, %xmm0, %xmm1
-; AVX-NEXT:    vpxor %xmm1, %xmm0, %xmm0
+; AVX-NEXT:    vpxor %xmm0, %xmm1, %xmm0
 ; AVX-NEXT:    vmovd %xmm0, %eax
 ; AVX-NEXT:    # kill: def $ax killed $ax killed $eax
 ; AVX-NEXT:    retq
diff --git a/llvm/test/CodeGen/X86/vector-trunc-packus.ll b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
--- a/llvm/test/CodeGen/X86/vector-trunc-packus.ll
+++ b/llvm/test/CodeGen/X86/vector-trunc-packus.ll
@@ -37,10 +37,9 @@
 ; SSE2-NEXT:    pxor %xmm1, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm2
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm1, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm4, %xmm0
+; SSE2-NEXT:    pand %xmm2, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm1
 ; SSE2-NEXT:    pand %xmm3, %xmm1
@@ -68,10 +67,9 @@
 ; SSSE3-NEXT:    pxor %xmm1, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm0, %xmm2
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm1, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm4, %xmm0
+; SSSE3-NEXT:    pand %xmm2, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm0, %xmm1
 ; SSSE3-NEXT:    pand %xmm3, %xmm1
@@ -186,10 +184,9 @@
 ; SSE2-NEXT:    pxor %xmm1, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm2
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm2
-; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm1, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm4, %xmm0
+; SSE2-NEXT:    pand %xmm2, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm1
 ; SSE2-NEXT:    pand %xmm3, %xmm1
@@ -218,10 +215,9 @@
 ; SSSE3-NEXT:    pxor %xmm1, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm0, %xmm2
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm2
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm1, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm4, %xmm0
+; SSSE3-NEXT:    pand %xmm2, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm0, %xmm1
 ; SSSE3-NEXT:    pand %xmm3, %xmm1
@@ -354,10 +350,9 @@
 ; SSE2-NEXT:    pxor %xmm2, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm1
 ; SSE2-NEXT:    pcmpgtd %xmm2, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm2, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm0
+; SSE2-NEXT:    pand %xmm1, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
 ; SSE2-NEXT:    por %xmm0, %xmm1
 ; SSE2-NEXT:    pand %xmm4, %xmm1
@@ -365,10 +360,9 @@
 ; SSE2-NEXT:    pxor %xmm2, %xmm0
 ; SSE2-NEXT:    movdqa %xmm0, %xmm4
 ; SSE2-NEXT:    pcmpgtd %xmm2, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
 ; SSE2-NEXT:    pcmpeqd %xmm2, %xmm0
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm2
+; SSE2-NEXT:    pand %xmm4, %xmm2
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
 ; SSE2-NEXT:    por %xmm2, %xmm0
 ; SSE2-NEXT:    pand %xmm3, %xmm0
@@ -410,10 +404,9 @@
 ; SSSE3-NEXT:    pxor %xmm2, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm0, %xmm1
 ; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm2, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm0
+; SSSE3-NEXT:    pand %xmm1, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm0, %xmm1
 ; SSSE3-NEXT:    pand %xmm4, %xmm1
@@ -421,10 +414,9 @@
 ; SSSE3-NEXT:    pxor %xmm2, %xmm0
 ; SSSE3-NEXT:    movdqa %xmm0, %xmm4
 ; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
 ; SSSE3-NEXT:    pcmpeqd %xmm2, %xmm0
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm2
+; SSSE3-NEXT:    pand %xmm4, %xmm2
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm2, %xmm0
 ; SSSE3-NEXT:    pand %xmm3, %xmm0
@@ -584,9 +576,9 @@
 ; SSE2-NEXT:    movdqa 32(%rdi), %xmm6
 ; SSE2-NEXT:    movdqa 48(%rdi), %xmm9
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm8 = [4294967295,4294967295]
-; SSE2-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648]
+; SSE2-NEXT:    movdqa {{.*#+}} xmm0 = [2147483648,2147483648]
 ; SSE2-NEXT:    movdqa %xmm3, %xmm2
-; SSE2-NEXT:    pxor %xmm11, %xmm2
+; SSE2-NEXT:    pxor %xmm0, %xmm2
 ; SSE2-NEXT:    movdqa {{.*#+}} xmm10 = [2147483647,2147483647]
 ; SSE2-NEXT:    movdqa %xmm10, %xmm5
 ; SSE2-NEXT:    pcmpgtd %xmm2, %xmm5
@@ -600,7 +592,7 @@
 ; SSE2-NEXT:    pandn %xmm8, %xmm2
 ; SSE2-NEXT:    por %xmm3, %xmm2
 ; SSE2-NEXT:    movdqa %xmm7, %xmm1
-; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    pxor %xmm0, %xmm1
 ; SSE2-NEXT:    movdqa %xmm10, %xmm3
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm3
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2]
@@ -613,7 +605,7 @@
 ; SSE2-NEXT:    pandn %xmm8, %xmm3
 ; SSE2-NEXT:    por %xmm7, %xmm3
 ; SSE2-NEXT:    movdqa %xmm6, %xmm1
-; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    pxor %xmm0, %xmm1
 ; SSE2-NEXT:    movdqa %xmm10, %xmm4
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm4
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
@@ -626,7 +618,7 @@
 ; SSE2-NEXT:    pandn %xmm8, %xmm7
 ; SSE2-NEXT:    por %xmm6, %xmm7
 ; SSE2-NEXT:    movdqa %xmm9, %xmm1
-; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    pxor %xmm0, %xmm1
 ; SSE2-NEXT:    movdqa %xmm10, %xmm4
 ; SSE2-NEXT:    pcmpgtd %xmm1, %xmm4
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
@@ -639,51 +631,47 @@
 ; SSE2-NEXT:    pandn %xmm8, %xmm4
 ; SSE2-NEXT:    por %xmm9, %xmm4
 ; SSE2-NEXT:    movdqa %xmm4, %xmm1
-; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    pxor %xmm0, %xmm1
 ; SSE2-NEXT:    movdqa %xmm1, %xmm5
-; SSE2-NEXT:    pcmpgtd %xmm11, %xmm5
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm11, %xmm1
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm5
+; SSE2-NEXT:    pcmpeqd %xmm0, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; SSE2-NEXT:    pand %xmm6, %xmm1
+; SSE2-NEXT:    pand %xmm5, %xmm1
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
 ; SSE2-NEXT:    por %xmm1, %xmm5
 ; SSE2-NEXT:    pand %xmm4, %xmm5
 ; SSE2-NEXT:    movdqa %xmm7, %xmm1
-; SSE2-NEXT:    pxor %xmm11, %xmm1
+; SSE2-NEXT:    pxor %xmm0, %xmm1
 ; SSE2-NEXT:    movdqa %xmm1, %xmm4
-; SSE2-NEXT:    pcmpgtd %xmm11, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm11, %xmm1
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3]
-; SSE2-NEXT:    pand %xmm6, %xmm0
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSE2-NEXT:    pcmpeqd %xmm0, %xmm1
+; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3]
+; SSE2-NEXT:    pand %xmm4, %xmm6
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm4[1,1,3,3]
-; SSE2-NEXT:    por %xmm0, %xmm1
+; SSE2-NEXT:    por %xmm6, %xmm1
 ; SSE2-NEXT:    pand %xmm7, %xmm1
 ; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,2],xmm5[0,2]
-; SSE2-NEXT:    movdqa %xmm3, %xmm0
-; SSE2-NEXT:    pxor %xmm11, %xmm0
-; SSE2-NEXT:    movdqa %xmm0, %xmm4
-; SSE2-NEXT:    pcmpgtd %xmm11, %xmm4
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm0
+; SSE2-NEXT:    movdqa %xmm3, %xmm4
+; SSE2-NEXT:    pxor %xmm0, %xmm4
+; SSE2-NEXT:    movdqa %xmm4, %xmm5
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm5
+; SSE2-NEXT:    pcmpeqd %xmm0, %xmm4
 ; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
-; SSE2-NEXT:    por %xmm0, %xmm4
-; SSE2-NEXT:    pand %xmm3, %xmm4
-; SSE2-NEXT:    movdqa %xmm2, %xmm0
-; SSE2-NEXT:    pxor %xmm11, %xmm0
-; SSE2-NEXT:    movdqa %xmm0, %xmm3
-; SSE2-NEXT:    pcmpgtd %xmm11, %xmm3
-; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm3[0,0,2,2]
-; SSE2-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3]
-; SSE2-NEXT:    pand %xmm5, %xmm6
-; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
-; SSE2-NEXT:    por %xmm6, %xmm0
+; SSE2-NEXT:    pand %xmm5, %xmm4
+; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
+; SSE2-NEXT:    por %xmm4, %xmm5
+; SSE2-NEXT:    pand %xmm3, %xmm5
+; SSE2-NEXT:    movdqa %xmm2, %xmm3
+; SSE2-NEXT:    pxor %xmm0, %xmm3
+; SSE2-NEXT:    movdqa %xmm3, %xmm4
+; SSE2-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSE2-NEXT:    pcmpeqd %xmm0, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; SSE2-NEXT:    pand %xmm4, %xmm3
+; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
+; SSE2-NEXT:    por %xmm3, %xmm0
 ; SSE2-NEXT:    pand %xmm2, %xmm0
-; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm4[0,2]
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm5[0,2]
 ; SSE2-NEXT:    retq
 ;
 ; SSSE3-LABEL: trunc_packus_v8i64_v8i32:
@@ -693,9 +681,9 @@
 ; SSSE3-NEXT:    movdqa 32(%rdi), %xmm6
 ; SSSE3-NEXT:    movdqa 48(%rdi), %xmm9
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm8 = [4294967295,4294967295]
-; SSSE3-NEXT:    movdqa {{.*#+}} xmm11 = [2147483648,2147483648]
+; SSSE3-NEXT:    movdqa {{.*#+}} xmm0 = [2147483648,2147483648]
 ; SSSE3-NEXT:    movdqa %xmm3, %xmm2
-; SSSE3-NEXT:    pxor %xmm11, %xmm2
+; SSSE3-NEXT:    pxor %xmm0, %xmm2
 ; SSSE3-NEXT:    movdqa {{.*#+}} xmm10 = [2147483647,2147483647]
 ; SSSE3-NEXT:    movdqa %xmm10, %xmm5
 ; SSSE3-NEXT:    pcmpgtd %xmm2, %xmm5
@@ -709,7 +697,7 @@
 ; SSSE3-NEXT:    pandn %xmm8, %xmm2
 ; SSSE3-NEXT:    por %xmm3, %xmm2
 ; SSSE3-NEXT:    movdqa %xmm7, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    pxor %xmm0, %xmm1
 ; SSSE3-NEXT:    movdqa %xmm10, %xmm3
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm3
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm3[0,0,2,2]
@@ -722,7 +710,7 @@
 ; SSSE3-NEXT:    pandn %xmm8, %xmm3
 ; SSSE3-NEXT:    por %xmm7, %xmm3
 ; SSSE3-NEXT:    movdqa %xmm6, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    pxor %xmm0, %xmm1
 ; SSSE3-NEXT:    movdqa %xmm10, %xmm4
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm4
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
@@ -735,7 +723,7 @@
 ; SSSE3-NEXT:    pandn %xmm8, %xmm7
 ; SSSE3-NEXT:    por %xmm6, %xmm7
 ; SSSE3-NEXT:    movdqa %xmm9, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    pxor %xmm0, %xmm1
 ; SSSE3-NEXT:    movdqa %xmm10, %xmm4
 ; SSSE3-NEXT:    pcmpgtd %xmm1, %xmm4
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
@@ -748,51 +736,47 @@
 ; SSSE3-NEXT:    pandn %xmm8, %xmm4
 ; SSSE3-NEXT:    por %xmm9, %xmm4
 ; SSSE3-NEXT:    movdqa %xmm4, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    pxor %xmm0, %xmm1
 ; SSSE3-NEXT:    movdqa %xmm1, %xmm5
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm5
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm1
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm5
+; SSSE3-NEXT:    pcmpeqd %xmm0, %xmm1
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm6, %xmm1
+; SSSE3-NEXT:    pand %xmm5, %xmm1
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
 ; SSSE3-NEXT:    por %xmm1, %xmm5
 ; SSSE3-NEXT:    pand %xmm4, %xmm5
 ; SSSE3-NEXT:    movdqa %xmm7, %xmm1
-; SSSE3-NEXT:    pxor %xmm11, %xmm1
+; SSSE3-NEXT:    pxor %xmm0, %xmm1
 ; SSSE3-NEXT:    movdqa %xmm1, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm1
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm6, %xmm0
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSSE3-NEXT:    pcmpeqd %xmm0, %xmm1
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm4, %xmm6
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm1 = xmm4[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm1
+; SSSE3-NEXT:    por %xmm6, %xmm1
 ; SSSE3-NEXT:    pand %xmm7, %xmm1
 ; SSSE3-NEXT:    shufps {{.*#+}} xmm1 = xmm1[0,2],xmm5[0,2]
-; SSSE3-NEXT:    movdqa %xmm3, %xmm0
-; SSSE3-NEXT:    pxor %xmm11, %xmm0
-; SSSE3-NEXT:    movdqa %xmm0, %xmm4
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm4
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm4[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm0
+; SSSE3-NEXT:    movdqa %xmm3, %xmm4
+; SSSE3-NEXT:    pxor %xmm0, %xmm4
+; SSSE3-NEXT:    movdqa %xmm4, %xmm5
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm5
+; SSSE3-NEXT:    pcmpeqd %xmm0, %xmm4
 ; SSSE3-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,1,3,3]
-; SSSE3-NEXT:    por %xmm0, %xmm4
-; SSSE3-NEXT:    pand %xmm3, %xmm4
-; SSSE3-NEXT:    movdqa %xmm2, %xmm0
-; SSSE3-NEXT:    pxor %xmm11, %xmm0
-; SSSE3-NEXT:    movdqa %xmm0, %xmm3
-; SSSE3-NEXT:    pcmpgtd %xmm11, %xmm3
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm3[0,0,2,2]
-; SSSE3-NEXT:    pcmpeqd %xmm11, %xmm0
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm6 = xmm0[1,1,3,3]
-; SSSE3-NEXT:    pand %xmm5, %xmm6
-; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,3,3]
-; SSSE3-NEXT:    por %xmm6, %xmm0
+; SSSE3-NEXT:    pand %xmm5, %xmm4
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[1,1,3,3]
+; SSSE3-NEXT:    por %xmm4, %xmm5
+; SSSE3-NEXT:    pand %xmm3, %xmm5
+; SSSE3-NEXT:    movdqa %xmm2, %xmm3
+; SSSE3-NEXT:    pxor %xmm0, %xmm3
+; SSSE3-NEXT:    movdqa %xmm3, %xmm4
+; SSSE3-NEXT:    pcmpgtd %xmm0, %xmm4
+; SSSE3-NEXT:    pcmpeqd %xmm0, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
+; SSSE3-NEXT:    pand %xmm4, %xmm3
+; SSSE3-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[1,1,3,3]
+; SSSE3-NEXT:    por %xmm3, %xmm0
 ; SSSE3-NEXT:    pand %xmm2, %xmm0
-; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm4[0,2]
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm5[0,2]
 ; SSSE3-NEXT:    retq
 ;
 ; SSE41-LABEL: trunc_packus_v8i64_v8i32:
diff --git a/llvm/test/CodeGen/X86/widen_cast-5.ll b/llvm/test/CodeGen/X86/widen_cast-5.ll
--- a/llvm/test/CodeGen/X86/widen_cast-5.ll
+++ b/llvm/test/CodeGen/X86/widen_cast-5.ll
@@ -15,9 +15,9 @@
 ;
 ; X64-LABEL: convert:
 ; X64:       ## %bb.0: ## %entry
-; X64-NEXT:    movq %rsi, %xmm0
-; X64-NEXT:    pxor {{.*}}(%rip), %xmm0
-; X64-NEXT:    movq %xmm0, (%rdi)
+; X64-NEXT:    movabsq $140733193388287, %rax ## imm = 0x7FFF000000FF
+; X64-NEXT:    xorq %rsi, %rax
+; X64-NEXT:    movq %rax, (%rdi)
 ; X64-NEXT:    retq
 entry:
 	%conv = bitcast i64 %src to <2 x i32>
diff --git a/llvm/test/CodeGen/X86/xor.ll b/llvm/test/CodeGen/X86/xor.ll
--- a/llvm/test/CodeGen/X86/xor.ll
+++ b/llvm/test/CodeGen/X86/xor.ll
@@ -394,14 +394,11 @@
 define i32 @PR17487(i1 %tobool) {
 ; X86-LABEL: PR17487:
 ; X86:       # %bb.0:
-; X86-NEXT:    movd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X86-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
-; X86-NEXT:    pandn {{\.LCPI.*}}, %xmm0
-; X86-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,3,2,3]
-; X86-NEXT:    movd %xmm0, %ecx
+; X86-NEXT:    movb {{[0-9]+}}(%esp), %cl
+; X86-NEXT:    notb %cl
 ; X86-NEXT:    xorl %eax, %eax
-; X86-NEXT:    cmpl $1, %ecx
-; X86-NEXT:    setne %al
+; X86-NEXT:    testb $1, %cl
+; X86-NEXT:    sete %al
 ; X86-NEXT:    retl
 ;
 ; X64-LIN-LABEL: PR17487: