Index: lib/Target/X86/X86ISelLowering.cpp
===================================================================
--- lib/Target/X86/X86ISelLowering.cpp
+++ lib/Target/X86/X86ISelLowering.cpp
@@ -19829,20 +19829,23 @@
 }
 
 static SDValue getGatherNode(unsigned Opc, SDValue Op, SelectionDAG &DAG,
-                              SDValue Src, SDValue Mask, SDValue Base,
-                              SDValue Index, SDValue ScaleOp, SDValue Chain,
-                              const X86Subtarget &Subtarget) {
+                             SDValue Src, SDValue Mask, SDValue Base,
+                             SDValue Index, SDValue ScaleOp, SDValue Chain,
+                             const X86Subtarget &Subtarget) {
   SDLoc dl(Op);
   auto *C = cast<ConstantSDNode>(ScaleOp);
   SDValue Scale = DAG.getTargetConstant(C->getZExtValue(), dl, MVT::i8);
   MVT MaskVT = MVT::getVectorVT(MVT::i1,
                              Index.getSimpleValueType().getVectorNumElements());
+  EVT VT = Op.getValueType();
 
   SDValue VMask = getMaskNode(Mask, MaskVT, Subtarget, DAG, dl);
-  SDVTList VTs = DAG.getVTList(Op.getValueType(), MaskVT, MVT::Other);
+  SDVTList VTs = DAG.getVTList(VT, MaskVT, MVT::Other);
   SDValue Disp = DAG.getTargetConstant(0, dl, MVT::i32);
   SDValue Segment = DAG.getRegister(0, MVT::i32);
-  if (Src.isUndef())
+  if (ISD::isBuildVectorAllOnes(VMask.getNode()))
+    Src = DAG.getUNDEF(VT); // If the mask is all ones we don't need a source.
+  else if (Src.isUndef())
     Src = getZeroVector(Op.getSimpleValueType(), Subtarget, DAG, dl);
   SDValue Ops[] = {Src, VMask, Base, Scale, Index, Disp, Segment, Chain};
   SDNode *Res = DAG.getMachineNode(Opc, dl, VTs, Ops);
Index: test/CodeGen/X86/avx512-gather-scatter-intrin.ll
===================================================================
--- test/CodeGen/X86/avx512-gather-scatter-intrin.ll
+++ test/CodeGen/X86/avx512-gather-scatter-intrin.ll
@@ -297,11 +297,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3div2_df:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovapd %xmm0, %xmm2
-; CHECK-NEXT:    vgatherqpd (%rdi,%xmm1,4), %xmm2 {%k1}
+; CHECK-NEXT:    vgatherqpd (%rdi,%xmm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherqpd (%rdi,%xmm1,2), %xmm0 {%k1}
-; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vgatherqpd (%rdi,%xmm1,2), %xmm2 {%k1}
+; CHECK-NEXT:    vaddpd %xmm2, %xmm0, %xmm0
 ; CHECK-NEXT:    retq
   %res = call <2 x double> @llvm.x86.avx512.gather3div2.df(<2 x double> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 4)
   %res1 = call <2 x double> @llvm.x86.avx512.gather3div2.df(<2 x double> %x0, i8* %x1, <2 x i64> %x2, i8 -1, i32 2)
@@ -330,11 +329,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3div4_df:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovapd %ymm0, %ymm2
-; CHECK-NEXT:    vgatherqpd (%rdi,%ymm1,4), %ymm2 {%k1}
+; CHECK-NEXT:    vgatherqpd (%rdi,%ymm1,4), %ymm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherqpd (%rdi,%ymm1,2), %ymm0 {%k1}
-; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vgatherqpd (%rdi,%ymm1,2), %ymm2 {%k1}
+; CHECK-NEXT:    vaddpd %ymm2, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
   %res = call <4 x double> @llvm.x86.avx512.gather3div4.df(<4 x double> %x0, i8* %x1, <4 x i64> %x2, i8 %x3, i32 4)
   %res1 = call <4 x double> @llvm.x86.avx512.gather3div4.df(<4 x double> %x0, i8* %x1, <4 x i64> %x2, i8 -1, i32 2)
@@ -348,11 +346,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3div4_di:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovdqa %ymm0, %ymm2
-; CHECK-NEXT:    vpgatherqq (%rdi,%ymm1,8), %ymm2 {%k1}
-; CHECK-NEXT:    kxnorw %k0, %k0, %k1
 ; CHECK-NEXT:    vpgatherqq (%rdi,%ymm1,8), %ymm0 {%k1}
-; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    kxnorw %k0, %k0, %k1
+; CHECK-NEXT:    vpgatherqq (%rdi,%ymm1,8), %ymm2 {%k1}
+; CHECK-NEXT:    vpaddq %ymm2, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
   %res = call <4 x i64> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 %x3, i32 8)
   %res1 = call <4 x i64> @llvm.x86.avx512.gather3div4.di(<4 x i64> %x0, i8* %x1, <4 x i64> %x2, i8 -1, i32 8)
@@ -366,11 +363,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3div4_sf:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovaps %xmm0, %xmm2
-; CHECK-NEXT:    vgatherqps (%rdi,%xmm1,4), %xmm2 {%k1}
+; CHECK-NEXT:    vgatherqps (%rdi,%xmm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherqps (%rdi,%xmm1,2), %xmm0 {%k1}
-; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vgatherqps (%rdi,%xmm1,2), %xmm2 {%k1}
+; CHECK-NEXT:    vaddps %xmm2, %xmm0, %xmm0
 ; CHECK-NEXT:    retq
   %res = call <4 x float> @llvm.x86.avx512.gather3div4.sf(<4 x float> %x0, i8* %x1, <2 x i64> %x2, i8 %x3, i32 4)
   %res1 = call <4 x float> @llvm.x86.avx512.gather3div4.sf(<4 x float> %x0, i8* %x1, <2 x i64> %x2, i8 -1, i32 2)
@@ -385,7 +381,6 @@
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k2
-; CHECK-NEXT:    vmovdqa %xmm0, %xmm2
 ; CHECK-NEXT:    vpgatherqd (%rdi,%xmm1,4), %xmm2 {%k2}
 ; CHECK-NEXT:    vpgatherqd (%rdi,%xmm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
@@ -402,11 +397,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3div8_sf:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovaps %xmm0, %xmm2
-; CHECK-NEXT:    vgatherqps (%rdi,%ymm1,4), %xmm2 {%k1}
+; CHECK-NEXT:    vgatherqps (%rdi,%ymm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherqps (%rdi,%ymm1,2), %xmm0 {%k1}
-; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vgatherqps (%rdi,%ymm1,2), %xmm2 {%k1}
+; CHECK-NEXT:    vaddps %xmm2, %xmm0, %xmm0
 ; CHECK-NEXT:    vzeroupper
 ; CHECK-NEXT:    retq
   %res = call <4 x float> @llvm.x86.avx512.gather3div8.sf(<4 x float> %x0, i8* %x1, <4 x i64> %x2, i8 %x3, i32 4)
@@ -440,11 +434,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3siv2_df:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovapd %xmm0, %xmm2
-; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,4), %xmm2 {%k1}
+; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,2), %xmm0 {%k1}
-; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,2), %xmm2 {%k1}
+; CHECK-NEXT:    vaddpd %xmm2, %xmm0, %xmm0
 ; CHECK-NEXT:    retq
   %res = call <2 x double> @llvm.x86.avx512.gather3siv2.df(<2 x double> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 4)
   %res1 = call <2 x double> @llvm.x86.avx512.gather3siv2.df(<2 x double> %x0, i8* %x1, <4 x i32> %x2, i8 -1, i32 2)
@@ -473,11 +466,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3siv4_df:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovapd %ymm0, %ymm2
-; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,4), %ymm2 {%k1}
+; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,4), %ymm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,2), %ymm0 {%k1}
-; CHECK-NEXT:    vaddpd %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vgatherdpd (%rdi,%xmm1,2), %ymm2 {%k1}
+; CHECK-NEXT:    vaddpd %ymm2, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
   %res = call <4 x double> @llvm.x86.avx512.gather3siv4.df(<4 x double> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 4)
   %res1 = call <4 x double> @llvm.x86.avx512.gather3siv4.df(<4 x double> %x0, i8* %x1, <4 x i32> %x2, i8 -1, i32 2)
@@ -506,11 +498,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3siv4_sf:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovaps %xmm0, %xmm2
-; CHECK-NEXT:    vgatherdps (%rdi,%xmm1,4), %xmm2 {%k1}
+; CHECK-NEXT:    vgatherdps (%rdi,%xmm1,4), %xmm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherdps (%rdi,%xmm1,2), %xmm0 {%k1}
-; CHECK-NEXT:    vaddps %xmm0, %xmm2, %xmm0
+; CHECK-NEXT:    vgatherdps (%rdi,%xmm1,2), %xmm2 {%k1}
+; CHECK-NEXT:    vaddps %xmm2, %xmm0, %xmm0
 ; CHECK-NEXT:    retq
   %res = call <4 x float> @llvm.x86.avx512.gather3siv4.sf(<4 x float> %x0, i8* %x1, <4 x i32> %x2, i8 %x3, i32 4)
   %res1 = call <4 x float> @llvm.x86.avx512.gather3siv4.sf(<4 x float> %x0, i8* %x1, <4 x i32> %x2, i8 -1, i32 2)
@@ -525,7 +516,6 @@
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k2
-; CHECK-NEXT:    vmovdqa %xmm0, %xmm2
 ; CHECK-NEXT:    vpgatherdd (%rdi,%xmm1,4), %xmm2 {%k2}
 ; CHECK-NEXT:    vpgatherdd (%rdi,%xmm1,2), %xmm0 {%k1}
 ; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0
@@ -542,11 +532,10 @@
 ; CHECK-LABEL: test_int_x86_avx512_gather3siv8_sf:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kmovb %esi, %k1
-; CHECK-NEXT:    vmovaps %ymm0, %ymm2
-; CHECK-NEXT:    vgatherdps (%rdi,%ymm1,4), %ymm2 {%k1}
+; CHECK-NEXT:    vgatherdps (%rdi,%ymm1,4), %ymm0 {%k1}
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vgatherdps (%rdi,%ymm1,2), %ymm0 {%k1}
-; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0
+; CHECK-NEXT:    vgatherdps (%rdi,%ymm1,2), %ymm2 {%k1}
+; CHECK-NEXT:    vaddps %ymm2, %ymm0, %ymm0
 ; CHECK-NEXT:    retq
   %res = call <8 x float> @llvm.x86.avx512.gather3siv8.sf(<8 x float> %x0, i8* %x1, <8 x i32> %x2, i8 %x3, i32 4)
   %res1 = call <8 x float> @llvm.x86.avx512.gather3siv8.sf(<8 x float> %x0, i8* %x1, <8 x i32> %x2, i8 -1, i32 2)
@@ -846,7 +835,6 @@
 ; CHECK-LABEL: gather_mask_test:
 ; CHECK:       ## BB#0:
 ; CHECK-NEXT:    kxnorw %k0, %k0, %k1
-; CHECK-NEXT:    vmovaps %zmm1, %zmm2
 ; CHECK-NEXT:    vgatherdps (%rdi,%zmm0,4), %zmm2 {%k1}
 ; CHECK-NEXT:    kxorw %k0, %k0, %k1
 ; CHECK-NEXT:    vmovaps %zmm1, %zmm3