diff --git a/llvm/lib/Target/DirectX/DXILPrepare.cpp b/llvm/lib/Target/DirectX/DXILPrepare.cpp
--- a/llvm/lib/Target/DirectX/DXILPrepare.cpp
+++ b/llvm/lib/Target/DirectX/DXILPrepare.cpp
@@ -15,6 +15,7 @@
 #include "DirectXIRPasses/PointerTypeAnalysis.h"
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringSet.h"
 #include "llvm/CodeGen/Passes.h"
 #include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/Instruction.h"
@@ -79,6 +80,55 @@
                       Attr);
 }
 
+static void collectDeadStringAttrs(AttributeMask &DeadAttrs, AttributeSet &&AS,
+                                   StringSet<> LiveKeys) {
+  for (auto &Attr : AS) {
+    if (!Attr.isStringAttribute())
+      continue;
+    StringRef Key = Attr.getKindAsString();
+    if (LiveKeys.contains(Key))
+      continue;
+    DeadAttrs.addAttribute(Key);
+  }
+}
+
+static void removeStringFunctionAttributes(Function &F) {
+  AttributeList Attrs = F.getAttributes();
+  StringSet<> LiveKeys = {"waveops-include-helper-lanes"
+                          "fp32-denorm-mode"};
+  // Collect DeadKeys in FnAttrs.
+  AttributeMask DeadAttrs;
+  collectDeadStringAttrs(DeadAttrs, Attrs.getFnAttrs(), LiveKeys);
+  collectDeadStringAttrs(DeadAttrs, Attrs.getRetAttrs(), LiveKeys);
+
+  F.removeFnAttrs(DeadAttrs);
+  F.removeRetAttrs(DeadAttrs);
+}
+
+static void cleanModuleFlags(Module &M) {
+  constexpr StringLiteral DeadKeys[] = {"frame-pointer"};
+  // Collect DeadKeys in ModuleFlags.
+  StringSet<> DeadKeySet;
+  for (auto &Key : DeadKeys) {
+    if (M.getModuleFlag(Key))
+      DeadKeySet.insert(Key);
+  }
+  if (DeadKeySet.empty())
+    return;
+
+  SmallVector<Module::ModuleFlagEntry, 8> ModuleFlags;
+  M.getModuleFlagsMetadata(ModuleFlags);
+  NamedMDNode *MDFlags = M.getModuleFlagsMetadata();
+  MDFlags->eraseFromParent();
+  // Add ModuleFlag which not dead.
+  for (auto &Flag : ModuleFlags) {
+    StringRef Key = Flag.Key->getString();
+    if (DeadKeySet.contains(Key))
+      continue;
+    M.addModuleFlag(Flag.Behavior, Key, Flag.Val);
+  }
+}
+
 class DXILPrepareModule : public ModulePass {
 
   static Value *maybeGenerateBitcast(IRBuilder<> &Builder,
@@ -112,6 +162,7 @@
     for (auto &F : M.functions()) {
       F.removeFnAttrs(AttrMask);
       F.removeRetAttrs(AttrMask);
+      removeStringFunctionAttributes(F);
       for (size_t Idx = 0, End = F.arg_size(); Idx < End; ++Idx)
         F.removeParamAttrs(Idx, AttrMask);
 
@@ -170,6 +221,8 @@
         }
       }
     }
+    // Remove flags not in llvm3.7.
+    cleanModuleFlags(M);
     return true;
   }
 
diff --git a/llvm/lib/Target/DirectX/DirectXTargetMachine.cpp b/llvm/lib/Target/DirectX/DirectXTargetMachine.cpp
--- a/llvm/lib/Target/DirectX/DirectXTargetMachine.cpp
+++ b/llvm/lib/Target/DirectX/DirectXTargetMachine.cpp
@@ -76,8 +76,8 @@
   FunctionPass *createTargetRegisterAllocator(bool) override { return nullptr; }
   void addCodeGenPrepare() override {
     addPass(createDXILOpLoweringLegacyPass());
-    addPass(createDXILPrepareModulePass());
     addPass(createDXILTranslateMetadataPass());
+    addPass(createDXILPrepareModulePass());
   }
 };
 
diff --git a/llvm/test/CodeGen/DirectX/strip-string-fn-attrs-module-flags.ll b/llvm/test/CodeGen/DirectX/strip-string-fn-attrs-module-flags.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/CodeGen/DirectX/strip-string-fn-attrs-module-flags.ll
@@ -0,0 +1,47 @@
+; RUN: opt -S -dxil-prepare < %s | FileCheck %s
+target datalayout = "e-m:e-p:32:32-i1:32-i8:8-i16:16-i32:32-i64:64-f16:16-f32:32-f64:64-n8:16:32:64"
+target triple = "dxil-unknown-shadermodel6.0-compute"
+
+; make sure string function attribute and module flags are clear for llvm3.7.
+
+; CHECK-LABEL:define void @CSMain()
+; CHECK:attributes #0 = { nounwind memory(readwrite, inaccessiblemem: read) }
+; CHECK:!llvm.module.flags = !{!0}
+; CHECK:!0 = !{i32 1, !"wchar_size", i32 4}
+
+%"class.hlsl::RWBuffer" = type { ptr }
+%dx.types.Handle = type { ptr }
+%dx.types.ResRet.i32 = type { i32, i32, i32, i32, i32 }
+
+@"?srcBuffer@@3V?$RWBuffer@T?$__vector@H$01@__clang@@@hlsl@@A" = local_unnamed_addr global %"class.hlsl::RWBuffer" zeroinitializer, align 4
+@"?dstBuffer@@3V?$RWBuffer@T?$__vector@H$01@__clang@@@hlsl@@A" = local_unnamed_addr global %"class.hlsl::RWBuffer" zeroinitializer, align 4
+
+; Function Attrs: mustprogress nofree nounwind willreturn memory(readwrite, inaccessiblemem: read)
+define void @CSMain() local_unnamed_addr #0 {
+entry:
+  %0 = call %dx.types.Handle @dx.op.createHandle(i32 57, i8 1, i32 1, i32 0, i1 false)
+  %1 = call %dx.types.Handle @dx.op.createHandle(i32 57, i8 1, i32 0, i32 1, i1 false)
+  %2 = call i32 @dx.op.threadId.i32(i32 93, i32 0)
+  %3 = call %dx.types.ResRet.i32 @dx.op.bufferLoad.i32(i32 68, %dx.types.Handle %1, i32 %2, i32 poison)
+  %4 = extractvalue %dx.types.ResRet.i32 %3, 0
+  %5 = extractvalue %dx.types.ResRet.i32 %3, 1
+  %add.i.i0 = add i32 %4, 10
+  %add.i.i1 = add i32 %5, 10
+  call void @dx.op.bufferStore.i32(i32 69, %dx.types.Handle %0, i32 %2, i32 poison, i32 %add.i.i0, i32 %add.i.i1, i32 %add.i.i0, i32 %add.i.i0, i8 15)
+  ret void
+}
+
+declare %dx.types.Handle @dx.op.createHandle(i32 %0, i8 %1, i32 %2, i32 %3, i1 %4)
+
+declare %dx.types.ResRet.i32 @dx.op.bufferLoad.i32(i32 %0, %dx.types.Handle %1, i32 %2, i32 %3)
+
+declare void @dx.op.bufferStore.i32(i32 %0, %dx.types.Handle %1, i32 %2, i32 %3, i32 %4, i32 %5, i32 %6, i32 %7, i8 %8)
+
+declare i32 @dx.op.threadId.i32(i32 %0, i32 %1)
+
+attributes #0 = { mustprogress nofree nounwind willreturn memory(readwrite, inaccessiblemem: read) "frame-pointer"="all" "hlsl.numthreads"="1024,1,1" "hlsl.shader"="compute" "no-trapping-math"="true" "stack-protector-buffer-size"="8" }
+
+!llvm.module.flags = !{!0, !1}
+
+!0 = !{i32 1, !"wchar_size", i32 4}
+!1 = !{i32 7, !"frame-pointer", i32 2}