Index: lib/CodeGen/SelectionDAG/TargetLowering.cpp
===================================================================
--- lib/CodeGen/SelectionDAG/TargetLowering.cpp
+++ lib/CodeGen/SelectionDAG/TargetLowering.cpp
@@ -1633,7 +1633,12 @@
     break;
   }
   case ISD::ADD:
-  case ISD::SUB: {
+  case ISD::SUB:
+  case ISD::FADD:
+  case ISD::FSUB:
+  case ISD::FMUL:
+  case ISD::FDIV:
+  case ISD::FREM: {
     APInt SrcUndef, SrcZero;
     if (SimplifyDemandedVectorElts(Op.getOperand(1), DemandedElts, SrcUndef,
                                    SrcZero, TLO, Depth + 1))
Index: test/CodeGen/X86/avx-schedule.ll
===================================================================
--- test/CodeGen/X86/avx-schedule.ll
+++ test/CodeGen/X86/avx-schedule.ll
@@ -502,56 +502,48 @@
 define <4 x double> @test_blendpd(<4 x double> %a0, <4 x double> %a1, <4 x double> *%a2) {
 ; GENERIC-LABEL: test_blendpd:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.50]
 ; GENERIC-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:1.00]
 ; GENERIC-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
 ;
 ; SANDY-LABEL: test_blendpd:
 ; SANDY:       # %bb.0:
-; SANDY-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.50]
 ; SANDY-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:1.00]
 ; SANDY-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; SANDY-NEXT:    retq # sched: [1:1.00]
 ;
 ; HASWELL-LABEL: test_blendpd:
 ; HASWELL:       # %bb.0:
-; HASWELL-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.33]
 ; HASWELL-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:1.00]
 ; HASWELL-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; HASWELL-NEXT:    retq # sched: [7:1.00]
 ;
 ; BROADWELL-LABEL: test_blendpd:
 ; BROADWELL:       # %bb.0:
-; BROADWELL-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.33]
 ; BROADWELL-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:1.00]
 ; BROADWELL-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [7:0.50]
 ; BROADWELL-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKYLAKE-LABEL: test_blendpd:
 ; SKYLAKE:       # %bb.0:
-; SKYLAKE-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.33]
 ; SKYLAKE-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [4:0.50]
 ; SKYLAKE-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; SKYLAKE-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKX-LABEL: test_blendpd:
 ; SKX:       # %bb.0:
-; SKX-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.33]
 ; SKX-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [4:0.50]
 ; SKX-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; SKX-NEXT:    retq # sched: [7:1.00]
 ;
 ; BTVER2-LABEL: test_blendpd:
 ; BTVER2:       # %bb.0:
-; BTVER2-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:1.00]
 ; BTVER2-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:2.00]
 ; BTVER2-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [6:2.00]
 ; BTVER2-NEXT:    retq # sched: [4:1.00]
 ;
 ; ZNVER1-LABEL: test_blendpd:
 ; ZNVER1:       # %bb.0:
-; ZNVER1-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] sched: [1:0.50]
 ; ZNVER1-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 # sched: [3:1.00]
 ; ZNVER1-NEXT:    vblendpd {{.*#+}} ymm0 = ymm0[0],mem[1,2],ymm0[3] sched: [8:0.50]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
Index: test/CodeGen/X86/avx512-hadd-hsub.ll
===================================================================
--- test/CodeGen/X86/avx512-hadd-hsub.ll
+++ test/CodeGen/X86/avx512-hadd-hsub.ll
@@ -178,16 +178,16 @@
 define <4 x double> @fadd_noundef_low(<8 x double> %x225, <8 x double> %x227) {
 ; KNL-LABEL: fadd_noundef_low:
 ; KNL:       # %bb.0:
-; KNL-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; KNL-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; KNL-NEXT:    vunpcklpd {{.*#+}} ymm2 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
+; KNL-NEXT:    vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
 ; KNL-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
 ; KNL-NEXT:    # kill: def $ymm0 killed $ymm0 killed $zmm0
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: fadd_noundef_low:
 ; SKX:       # %bb.0:
-; SKX-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; SKX-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; SKX-NEXT:    vunpcklpd {{.*#+}} ymm2 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
+; SKX-NEXT:    vunpckhpd {{.*#+}} ymm0 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
 ; SKX-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
 ; SKX-NEXT:    # kill: def $ymm0 killed $ymm0 killed $zmm0
 ; SKX-NEXT:    retq
@@ -252,17 +252,15 @@
 define double @fadd_noundef_eel(<8 x double> %x225, <8 x double> %x227) {
 ; KNL-LABEL: fadd_noundef_eel:
 ; KNL:       # %bb.0:
-; KNL-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; KNL-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; KNL-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
+; KNL-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
+; KNL-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
 ; KNL-NEXT:    # kill: def $xmm0 killed $xmm0 killed $zmm0
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: fadd_noundef_eel:
 ; SKX:       # %bb.0:
-; SKX-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; SKX-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; SKX-NEXT:    vaddpd %zmm0, %zmm2, %zmm0
+; SKX-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
+; SKX-NEXT:    vaddpd %zmm1, %zmm0, %zmm0
 ; SKX-NEXT:    # kill: def $xmm0 killed $xmm0 killed $zmm0
 ; SKX-NEXT:    vzeroupper
 ; SKX-NEXT:    retq
@@ -278,18 +276,18 @@
 define double @fsub_noundef_ee (<8 x double> %x225, <8 x double> %x227) {
 ; KNL-LABEL: fsub_noundef_ee:
 ; KNL:       # %bb.0:
-; KNL-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; KNL-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; KNL-NEXT:    vsubpd %zmm0, %zmm2, %zmm0
+; KNL-NEXT:    vextractf32x4 $2, %zmm1, %xmm0
+; KNL-NEXT:    vbroadcastsd %xmm0, %zmm0
+; KNL-NEXT:    vsubpd %zmm1, %zmm0, %zmm0
 ; KNL-NEXT:    vextractf32x4 $2, %zmm0, %xmm0
 ; KNL-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; KNL-NEXT:    retq
 ;
 ; SKX-LABEL: fsub_noundef_ee:
 ; SKX:       # %bb.0:
-; SKX-NEXT:    vunpcklpd {{.*#+}} zmm2 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; SKX-NEXT:    vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; SKX-NEXT:    vsubpd %zmm0, %zmm2, %zmm0
+; SKX-NEXT:    vextractf32x4 $2, %zmm1, %xmm0
+; SKX-NEXT:    vbroadcastsd %xmm0, %zmm0
+; SKX-NEXT:    vsubpd %zmm1, %zmm0, %zmm0
 ; SKX-NEXT:    vextractf32x4 $2, %zmm0, %xmm0
 ; SKX-NEXT:    vpermilpd {{.*#+}} xmm0 = xmm0[1,0]
 ; SKX-NEXT:    vzeroupper
Index: test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
===================================================================
--- test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
+++ test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
@@ -7210,8 +7210,7 @@
 ; X86-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
 ; X86-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; X86-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; X86-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
+; X86-NEXT:    vhaddpd %xmm0, %xmm0, %xmm0
 ; X86-NEXT:    vmovlpd %xmm0, (%esp)
 ; X86-NEXT:    fldl (%esp)
 ; X86-NEXT:    movl %ebp, %esp
@@ -7226,8 +7225,7 @@
 ; X64-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
 ; X64-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; X64-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; X64-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
+; X64-NEXT:    vhaddpd %xmm0, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
 entry:
@@ -7301,8 +7299,7 @@
 ; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
-; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; X86-NEXT:    vhaddps %xmm0, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
 ; X86-NEXT:    popl %eax
@@ -7318,8 +7315,7 @@
 ; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
-; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; X64-NEXT:    vhaddps %xmm0, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
 entry:
@@ -7351,7 +7347,7 @@
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -7368,7 +7364,7 @@
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
@@ -7407,8 +7403,7 @@
 ; X86-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
 ; X86-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; X86-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; X86-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
+; X86-NEXT:    vhaddpd %xmm0, %xmm0, %xmm0
 ; X86-NEXT:    vmovlpd %xmm0, (%esp)
 ; X86-NEXT:    fldl (%esp)
 ; X86-NEXT:    movl %ebp, %esp
@@ -7425,8 +7420,7 @@
 ; X64-NEXT:    vaddpd %ymm1, %ymm0, %ymm0
 ; X64-NEXT:    vextractf128 $1, %ymm0, %xmm1
 ; X64-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
-; X64-NEXT:    vaddpd %xmm1, %xmm0, %xmm0
+; X64-NEXT:    vhaddpd %xmm0, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
 entry:
@@ -7513,8 +7507,7 @@
 ; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
-; X86-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; X86-NEXT:    vhaddps %xmm0, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
 ; X86-NEXT:    popl %eax
@@ -7532,8 +7525,7 @@
 ; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
-; X64-NEXT:    vaddps %xmm1, %xmm0, %xmm0
+; X64-NEXT:    vhaddps %xmm0, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
 entry:
@@ -7570,7 +7562,7 @@
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
-; X86-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X86-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X86-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X86-NEXT:    vmovss %xmm0, (%esp)
 ; X86-NEXT:    flds (%esp)
@@ -7590,7 +7582,7 @@
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vpermilpd {{.*#+}} xmm1 = xmm0[1,0]
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
-; X64-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,0,3,2]
+; X64-NEXT:    vmovshdup {{.*#+}} xmm1 = xmm0[1,1,3,3]
 ; X64-NEXT:    vmulps %xmm1, %xmm0, %xmm0
 ; X64-NEXT:    vzeroupper
 ; X64-NEXT:    retq
Index: test/CodeGen/X86/sse41-schedule.ll
===================================================================
--- test/CodeGen/X86/sse41-schedule.ll
+++ test/CodeGen/X86/sse41-schedule.ll
@@ -21,112 +21,96 @@
 define <2 x double> @test_blendpd(<2 x double> %a0, <2 x double> %a1, <2 x double> *%a2) {
 ; GENERIC-LABEL: test_blendpd:
 ; GENERIC:       # %bb.0:
-; GENERIC-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; GENERIC-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; GENERIC-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; GENERIC-NEXT:    retq # sched: [1:1.00]
 ;
 ; SLM-LABEL: test_blendpd:
 ; SLM:       # %bb.0:
-; SLM-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:1.00]
 ; SLM-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; SLM-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [4:1.00]
 ; SLM-NEXT:    retq # sched: [4:1.00]
 ;
 ; SANDY-SSE-LABEL: test_blendpd:
 ; SANDY-SSE:       # %bb.0:
-; SANDY-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; SANDY-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; SANDY-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SANDY-SSE-NEXT:    retq # sched: [1:1.00]
 ;
 ; SANDY-LABEL: test_blendpd:
 ; SANDY:       # %bb.0:
-; SANDY-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; SANDY-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
 ; SANDY-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SANDY-NEXT:    retq # sched: [1:1.00]
 ;
 ; HASWELL-SSE-LABEL: test_blendpd:
 ; HASWELL-SSE:       # %bb.0:
-; HASWELL-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; HASWELL-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; HASWELL-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; HASWELL-SSE-NEXT:    retq # sched: [7:1.00]
 ;
 ; HASWELL-LABEL: test_blendpd:
 ; HASWELL:       # %bb.0:
-; HASWELL-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; HASWELL-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
 ; HASWELL-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; HASWELL-NEXT:    retq # sched: [7:1.00]
 ;
 ; BROADWELL-SSE-LABEL: test_blendpd:
 ; BROADWELL-SSE:       # %bb.0:
-; BROADWELL-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; BROADWELL-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; BROADWELL-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [6:0.50]
 ; BROADWELL-SSE-NEXT:    retq # sched: [7:1.00]
 ;
 ; BROADWELL-LABEL: test_blendpd:
 ; BROADWELL:       # %bb.0:
-; BROADWELL-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; BROADWELL-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
 ; BROADWELL-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [6:0.50]
 ; BROADWELL-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKYLAKE-SSE-LABEL: test_blendpd:
 ; SKYLAKE-SSE:       # %bb.0:
-; SKYLAKE-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; SKYLAKE-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [4:0.50]
 ; SKYLAKE-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SKYLAKE-SSE-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKYLAKE-LABEL: test_blendpd:
 ; SKYLAKE:       # %bb.0:
-; SKYLAKE-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; SKYLAKE-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [4:0.50]
 ; SKYLAKE-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SKYLAKE-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKX-SSE-LABEL: test_blendpd:
 ; SKX-SSE:       # %bb.0:
-; SKX-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; SKX-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [4:0.50]
 ; SKX-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SKX-SSE-NEXT:    retq # sched: [7:1.00]
 ;
 ; SKX-LABEL: test_blendpd:
 ; SKX:       # %bb.0:
-; SKX-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.33]
 ; SKX-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [4:0.50]
 ; SKX-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [7:0.50]
 ; SKX-NEXT:    retq # sched: [7:1.00]
 ;
 ; BTVER2-SSE-LABEL: test_blendpd:
 ; BTVER2-SSE:       # %bb.0:
-; BTVER2-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; BTVER2-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; BTVER2-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [6:1.00]
 ; BTVER2-SSE-NEXT:    retq # sched: [4:1.00]
 ;
 ; BTVER2-LABEL: test_blendpd:
 ; BTVER2:       # %bb.0:
-; BTVER2-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; BTVER2-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
 ; BTVER2-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [6:1.00]
 ; BTVER2-NEXT:    retq # sched: [4:1.00]
 ;
 ; ZNVER1-SSE-LABEL: test_blendpd:
 ; ZNVER1-SSE:       # %bb.0:
-; ZNVER1-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; ZNVER1-SSE-NEXT:    addpd %xmm1, %xmm0 # sched: [3:1.00]
 ; ZNVER1-SSE-NEXT:    blendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [8:0.50]
 ; ZNVER1-SSE-NEXT:    retq # sched: [1:0.50]
 ;
 ; ZNVER1-LABEL: test_blendpd:
 ; ZNVER1:       # %bb.0:
-; ZNVER1-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],xmm1[1] sched: [1:0.50]
 ; ZNVER1-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 # sched: [3:1.00]
 ; ZNVER1-NEXT:    vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1] sched: [8:0.50]
 ; ZNVER1-NEXT:    retq # sched: [1:0.50]
Index: test/CodeGen/X86/vector-shuffle-combining.ll
===================================================================
--- test/CodeGen/X86/vector-shuffle-combining.ll
+++ test/CodeGen/X86/vector-shuffle-combining.ll
@@ -2700,21 +2700,36 @@
 }
 
 define <4 x float> @PR22377(<4 x float> %a, <4 x float> %b) {
-; SSE-LABEL: PR22377:
-; SSE:       # %bb.0: # %entry
-; SSE-NEXT:    movaps %xmm0, %xmm1
-; SSE-NEXT:    shufps {{.*#+}} xmm1 = xmm1[1,3],xmm0[1,3]
-; SSE-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2,0,2]
-; SSE-NEXT:    addps %xmm0, %xmm1
-; SSE-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
-; SSE-NEXT:    retq
+; SSE2-LABEL: PR22377:
+; SSE2:       # %bb.0: # %entry
+; SSE2-NEXT:    movaps %xmm0, %xmm1
+; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[1,3],xmm0[2,3]
+; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2,0,2]
+; SSE2-NEXT:    addps %xmm0, %xmm1
+; SSE2-NEXT:    unpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; SSE2-NEXT:    retq
+;
+; SSSE3-LABEL: PR22377:
+; SSSE3:       # %bb.0: # %entry
+; SSSE3-NEXT:    movaps %xmm0, %xmm1
+; SSSE3-NEXT:    haddps %xmm0, %xmm1
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,1]
+; SSSE3-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSSE3-NEXT:    retq
+;
+; SSE41-LABEL: PR22377:
+; SSE41:       # %bb.0: # %entry
+; SSE41-NEXT:    movaps %xmm0, %xmm1
+; SSE41-NEXT:    haddps %xmm0, %xmm1
+; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,1]
+; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2,1,3]
+; SSE41-NEXT:    retq
 ;
 ; AVX-LABEL: PR22377:
 ; AVX:       # %bb.0: # %entry
-; AVX-NEXT:    vpermilps {{.*#+}} xmm1 = xmm0[1,3,1,3]
-; AVX-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,2,0,2]
-; AVX-NEXT:    vaddps %xmm0, %xmm1, %xmm1
-; AVX-NEXT:    vunpcklps {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; AVX-NEXT:    vhaddps %xmm0, %xmm0, %xmm1
+; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[0,2],xmm1[0,1]
+; AVX-NEXT:    vpermilps {{.*#+}} xmm0 = xmm0[0,2,1,3]
 ; AVX-NEXT:    retq
 entry:
   %s1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> <i32 1, i32 3, i32 1, i32 3>