diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index df5e16c6a54b..62577162f1d4 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -1575,38 +1575,32 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
     }
   }
 
-  // This block control legalization of v32i1/v64i1 which are available with
+  // This block control legalization of v32i1 which is available with
   // AVX512BW. 512-bit v32i16 and v64i8 vector legalization is controlled with
-  // useBWIRegs.
+  // useBWIRegs. v64i1 is also controled with useBWIRegs.
   if (!Subtarget.useSoftFloat() && Subtarget.hasBWI()) {
     addRegisterClass(MVT::v32i1,  &X86::VK32RegClass);
-    addRegisterClass(MVT::v64i1,  &X86::VK64RegClass);
 
-    for (auto VT : { MVT::v32i1, MVT::v64i1 }) {
-      setOperationAction(ISD::ADD,                VT, Custom);
-      setOperationAction(ISD::SUB,                VT, Custom);
-      setOperationAction(ISD::MUL,                VT, Custom);
-      setOperationAction(ISD::VSELECT,            VT, Expand);
-      setOperationAction(ISD::UADDSAT,            VT, Custom);
-      setOperationAction(ISD::SADDSAT,            VT, Custom);
-      setOperationAction(ISD::USUBSAT,            VT, Custom);
-      setOperationAction(ISD::SSUBSAT,            VT, Custom);
+    setOperationAction(ISD::ADD,                MVT::v32i1, Custom);
+    setOperationAction(ISD::SUB,                MVT::v32i1, Custom);
+    setOperationAction(ISD::MUL,                MVT::v32i1, Custom);
+    setOperationAction(ISD::VSELECT,            MVT::v32i1, Expand);
+    setOperationAction(ISD::UADDSAT,            MVT::v32i1, Custom);
+    setOperationAction(ISD::SADDSAT,            MVT::v32i1, Custom);
+    setOperationAction(ISD::USUBSAT,            MVT::v32i1, Custom);
+    setOperationAction(ISD::SSUBSAT,            MVT::v32i1, Custom);
 
-      setOperationAction(ISD::TRUNCATE,           VT, Custom);
-      setOperationAction(ISD::SETCC,              VT, Custom);
-      setOperationAction(ISD::EXTRACT_VECTOR_ELT, VT, Custom);
-      setOperationAction(ISD::INSERT_VECTOR_ELT,  VT, Custom);
-      setOperationAction(ISD::SELECT,             VT, Custom);
-      setOperationAction(ISD::BUILD_VECTOR,       VT, Custom);
-      setOperationAction(ISD::VECTOR_SHUFFLE,     VT, Custom);
-    }
+    setOperationAction(ISD::TRUNCATE,           MVT::v32i1, Custom);
+    setOperationAction(ISD::SETCC,              MVT::v32i1, Custom);
+    setOperationAction(ISD::EXTRACT_VECTOR_ELT, MVT::v32i1, Custom);
+    setOperationAction(ISD::INSERT_VECTOR_ELT,  MVT::v32i1, Custom);
+    setOperationAction(ISD::SELECT,             MVT::v32i1, Custom);
+    setOperationAction(ISD::BUILD_VECTOR,       MVT::v32i1, Custom);
+    setOperationAction(ISD::VECTOR_SHUFFLE,     MVT::v32i1, Custom);
 
     setOperationAction(ISD::CONCAT_VECTORS,     MVT::v32i1, Custom);
-    setOperationAction(ISD::CONCAT_VECTORS,     MVT::v64i1, Custom);
     setOperationAction(ISD::INSERT_SUBVECTOR,   MVT::v32i1, Custom);
-    setOperationAction(ISD::INSERT_SUBVECTOR,   MVT::v64i1, Custom);
-    for (auto VT : { MVT::v16i1, MVT::v32i1 })
-      setOperationAction(ISD::EXTRACT_SUBVECTOR, VT, Custom);
+    setOperationAction(ISD::EXTRACT_SUBVECTOR,  MVT::v16i1, Custom);
 
     // Extends from v32i1 masks to 256-bit vectors.
     setOperationAction(ISD::SIGN_EXTEND,        MVT::v32i8, Custom);
@@ -1696,6 +1690,34 @@ X86TargetLowering::X86TargetLowering(const X86TargetMachine &TM,
       setOperationAction(ISD::FSHL, MVT::v32i16, Custom);
       setOperationAction(ISD::FSHR, MVT::v32i16, Custom);
     }
+
+    // Only support v64i1 if we support v64i8. Without 64i8 we won't have any
+    // operations that can produce these values other than concatenating
+    // v32i1 vectors together. And we don't have any masked operations that
+    // need a v64i1. By making it legal we avoid needing to lower arbitrary
+    // shuffles of v64i1 which need v64i8 to be legal.
+    addRegisterClass(MVT::v64i1,  &X86::VK64RegClass);
+
+    setOperationAction(ISD::ADD,                MVT::v64i1, Custom);
+    setOperationAction(ISD::SUB,                MVT::v64i1, Custom);
+    setOperationAction(ISD::MUL,                MVT::v64i1, Custom);
+    setOperationAction(ISD::VSELECT,            MVT::v64i1, Expand);
+    setOperationAction(ISD::UADDSAT,            MVT::v64i1, Custom);
+    setOperationAction(ISD::SADDSAT,            MVT::v64i1, Custom);
+    setOperationAction(ISD::USUBSAT,            MVT::v64i1, Custom);
+    setOperationAction(ISD::SSUBSAT,            MVT::v64i1, Custom);
+
+    setOperationAction(ISD::TRUNCATE,           MVT::v64i1, Custom);
+    setOperationAction(ISD::SETCC,              MVT::v64i1, Custom);
+    setOperationAction(ISD::EXTRACT_VECTOR_ELT, MVT::v64i1, Custom);
+    setOperationAction(ISD::INSERT_VECTOR_ELT,  MVT::v64i1, Custom);
+    setOperationAction(ISD::SELECT,             MVT::v64i1, Custom);
+    setOperationAction(ISD::BUILD_VECTOR,       MVT::v64i1, Custom);
+    setOperationAction(ISD::VECTOR_SHUFFLE,     MVT::v64i1, Custom);
+
+    setOperationAction(ISD::CONCAT_VECTORS,     MVT::v64i1, Custom);
+    setOperationAction(ISD::INSERT_SUBVECTOR,   MVT::v64i1, Custom);
+    setOperationAction(ISD::EXTRACT_SUBVECTOR,  MVT::v32i1, Custom);
   }
 
   if (!Subtarget.useSoftFloat() && Subtarget.hasBWI()) {
diff --git a/llvm/test/CodeGen/X86/min-legal-vector-width.ll b/llvm/test/CodeGen/X86/min-legal-vector-width.ll
index 8bfd448f201c..6e256c060d29 100644
--- a/llvm/test/CodeGen/X86/min-legal-vector-width.ll
+++ b/llvm/test/CodeGen/X86/min-legal-vector-width.ll
@@ -1116,3 +1116,38 @@ define void @trunc_packus_v16i32_v16i8_store(<16 x i32>* %p, <16 x i8>* %q) "min
   store <16 x i8> %f, <16 x i8>* %q
   ret void
 }
+
+define <64 x i1> @v64i1_argument_return(<64 x i1> %x) "min-legal-vector-width"="256" {
+; CHECK-LABEL: v64i1_argument_return:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    retq
+  ret <64 x i1> %x
+}
+
+define void @v64i1_shuffle(<64 x i8>* %x, <64 x i8>* %y) "min-legal-vector-width"="256" {
+; CHECK-LABEL: v64i1_shuffle:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vmovdqa (%rdi), %ymm0
+; CHECK-NEXT:    vmovdqa 32(%rdi), %ymm1
+; CHECK-NEXT:    vptestnmb %ymm1, %ymm1, %k0
+; CHECK-NEXT:    vptestnmb %ymm0, %ymm0, %k1
+; CHECK-NEXT:    vpmovm2b %k1, %ymm2
+; CHECK-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14,1,0,3,2,5,4,7,6,9,8,11,10,13,12,15,14]
+; CHECK-NEXT:    vpshufb %ymm3, %ymm2, %ymm2
+; CHECK-NEXT:    vpmovb2m %ymm2, %k1
+; CHECK-NEXT:    vpmovm2b %k0, %ymm2
+; CHECK-NEXT:    vpshufb %ymm3, %ymm2, %ymm2
+; CHECK-NEXT:    vpmovb2m %ymm2, %k2
+; CHECK-NEXT:    vmovdqu8 %ymm1, 32(%rsi) {%k2}
+; CHECK-NEXT:    vmovdqu8 %ymm0, (%rsi) {%k1}
+; CHECK-NEXT:    vzeroupper
+; CHECK-NEXT:    retq
+entry:
+  %a = load <64 x i8>, <64 x i8>* %x
+  %b = icmp eq <64 x i8> %a, zeroinitializer
+  %shuf = shufflevector <64 x i1> %b, <64 x i1> undef, <64 x i32> <i32 1, i32 0, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6, i32 9, i32 8, i32 11, i32 10, i32 13, i32 12, i32 15, i32 14, i32 17, i32 16, i32 19, i32 18, i32 21, i32 20, i32 23, i32 22, i32 25, i32 24, i32 27, i32 26, i32 29, i32 28, i32 31, i32 30, i32 33, i32 32, i32 35, i32 34, i32 37, i32 36, i32 39, i32 38, i32 41, i32 40, i32 43, i32 42, i32 45, i32 44, i32 47, i32 46, i32 49, i32 48, i32 51, i32 50, i32 53, i32 52, i32 55, i32 54, i32 57, i32 56, i32 59, i32 58, i32 61, i32 60, i32 63, i32 62>
+  call void @llvm.masked.store.v64i8.p0v64i8(<64 x i8> %a, <64 x i8>* %y, i32 1, <64 x i1> %shuf)
+  ret void
+}
+declare void @llvm.masked.store.v64i8.p0v64i8(<64 x i8>, <64 x i8>*, i32, <64 x i1>)
+