diff --git a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
--- a/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
+++ b/llvm/lib/Transforms/Instrumentation/MemorySanitizer.cpp
@@ -1056,7 +1056,7 @@
   ValueMap<Value*, Value*> ShadowMap, OriginMap;
   std::unique_ptr<VarArgHelper> VAHelper;
   const TargetLibraryInfo *TLI;
-  Instruction *ActualFnStart;
+  Instruction *FnPrologueEnd;
 
   // The following flags disable parts of MSan instrumentation based on
   // exclusion list contents and command-line options.
@@ -1095,11 +1095,12 @@
     removeUnreachableBlocks(F);
 
     MS.initializeCallbacks(*F.getParent());
-    ActualFnStart = F.getEntryBlock().getFirstNonPHI();
+    FnPrologueEnd = IRBuilder<>(F.getEntryBlock().getFirstNonPHI())
+                        .CreateIntrinsic(Intrinsic::donothing, {}, {});
 
     if (MS.CompileKernel) {
-      IRBuilder<> IRB(ActualFnStart);
-      insertKmsanPrologue(IRB);
+      IRBuilder<> IRB(FnPrologueEnd);
+      FnPrologueEnd = insertKmsanPrologue(IRB);
     }
 
     LLVM_DEBUG(if (!InsertChecks) dbgs()
@@ -1107,6 +1108,11 @@
                << F.getName() << "'\n");
   }
 
+  bool isInPrologue(Instruction &I) {
+    return I.getParent() == FnPrologueEnd->getParent() &&
+           (&I == FnPrologueEnd || I.comesBefore(FnPrologueEnd));
+  }
+
   Value *updateOrigin(Value *V, IRBuilder<> &IRB) {
     if (MS.TrackOrigins <= 1) return V;
     return IRB.CreateCall(MS.MsanChainOriginFn, V);
@@ -1269,7 +1275,8 @@
     LLVM_DEBUG(dbgs() << "DONE:\n" << F);
   }
 
-  void insertKmsanPrologue(IRBuilder<> &IRB) {
+  // Returns the last instruction in the new prologue
+  Instruction *insertKmsanPrologue(IRBuilder<> &IRB) {
     Value *ContextState = IRB.CreateCall(MS.MsanGetContextStateFn, {});
     Constant *Zero = IRB.getInt32(0);
     MS.ParamTLS = IRB.CreateGEP(MS.MsanContextStateTy, ContextState,
@@ -1288,6 +1295,8 @@
     MS.RetvalOriginTLS =
         IRB.CreateGEP(MS.MsanContextStateTy, ContextState,
                       {Zero, IRB.getInt32(6)}, "retval_origin");
+
+    return cast<Instruction>(MS.RetvalOriginTLS);
   }
 
   /// Add MemorySanitizer instrumentation to a function.
@@ -1295,7 +1304,7 @@
     // Iterate all BBs in depth-first order and create shadow instructions
     // for all instructions (where applicable).
     // For PHI nodes we create dummy shadow PHIs which will be finalized later.
-    for (BasicBlock *BB : depth_first(ActualFnStart->getParent()))
+    for (BasicBlock *BB : depth_first(FnPrologueEnd->getParent()))
       visit(*BB);
 
     // Finalize PHI nodes.
@@ -1662,7 +1671,7 @@
       if (*ShadowPtr)
         return *ShadowPtr;
       Function *F = A->getParent();
-      IRBuilder<> EntryIRB(ActualFnStart);
+      IRBuilder<> EntryIRB(FnPrologueEnd);
       unsigned ArgOffset = 0;
       const DataLayout &DL = F->getParent()->getDataLayout();
       for (auto &FArg : F->args()) {
@@ -1880,9 +1889,8 @@
   void visit(Instruction &I) {
     if (I.getMetadata("nosanitize"))
       return;
-    // Don't want to visit if we're in the zone before ActualFnStart
-    if (I.getParent() == ActualFnStart->getParent() &&
-        I.comesBefore(ActualFnStart))
+    // Don't want to visit if we're in the prologue
+    if (isInPrologue(I))
       return;
     InstVisitor<MemorySanitizerVisitor>::visit(I);
   }
@@ -4309,7 +4317,7 @@
     if (!VAStartInstrumentationList.empty()) {
       // If there is a va_start in this function, make a backup copy of
       // va_arg_tls somewhere in the function entry block.
-      IRBuilder<> IRB(MSV.ActualFnStart);
+      IRBuilder<> IRB(MSV.FnPrologueEnd);
       VAArgOverflowSize =
           IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
       Value *CopySize =
@@ -4455,7 +4463,7 @@
   void finalizeInstrumentation() override {
     assert(!VAArgSize && !VAArgTLSCopy &&
            "finalizeInstrumentation called twice");
-    IRBuilder<> IRB(MSV.ActualFnStart);
+    IRBuilder<> IRB(MSV.FnPrologueEnd);
     VAArgSize = IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
     Value *CopySize = IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, 0),
                                     VAArgSize);
@@ -4648,7 +4656,7 @@
     if (!VAStartInstrumentationList.empty()) {
       // If there is a va_start in this function, make a backup copy of
       // va_arg_tls somewhere in the function entry block.
-      IRBuilder<> IRB(MSV.ActualFnStart);
+      IRBuilder<> IRB(MSV.FnPrologueEnd);
       VAArgOverflowSize =
           IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
       Value *CopySize =
@@ -4893,7 +4901,7 @@
   void finalizeInstrumentation() override {
     assert(!VAArgSize && !VAArgTLSCopy &&
            "finalizeInstrumentation called twice");
-    IRBuilder<> IRB(MSV.ActualFnStart);
+    IRBuilder<> IRB(MSV.FnPrologueEnd);
     VAArgSize = IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
     Value *CopySize = IRB.CreateAdd(ConstantInt::get(MS.IntptrTy, 0),
                                     VAArgSize);
@@ -5212,7 +5220,7 @@
     if (!VAStartInstrumentationList.empty()) {
       // If there is a va_start in this function, make a backup copy of
       // va_arg_tls somewhere in the function entry block.
-      IRBuilder<> IRB(MSV.ActualFnStart);
+      IRBuilder<> IRB(MSV.FnPrologueEnd);
       VAArgOverflowSize =
           IRB.CreateLoad(IRB.getInt64Ty(), MS.VAArgOverflowSizeTLS);
       Value *CopySize =
diff --git a/llvm/test/Instrumentation/MemorySanitizer/array_types.ll b/llvm/test/Instrumentation/MemorySanitizer/array_types.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/array_types.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/array_types.ll
@@ -18,8 +18,8 @@
 
 ; CHECK-LABEL: @InsertValue(
 ; CHECK-DAG: [[Sx:%.*]] = load i32, i32* {{.*}}@__msan_param_tls to i32*)
-; CHECK: [[A:%.*]] = insertvalue [2 x i32] [i32 -1, i32 -1], i32 [[Sx]], 0
 ; CHECK-DAG: [[Sy:%.*]] = load i32, i32* {{.*}}@__msan_param_tls to i64), i64 8) to i32*)
+; CHECK: [[A:%.*]] = insertvalue [2 x i32] [i32 -1, i32 -1], i32 [[Sx]], 0
 ; CHECK: [[B:%.*]] = insertvalue [2 x i32] [[A]], i32 [[Sy]], 1
 ; CHECK: store [2 x i32] [[B]], [2 x i32]* {{.*}}@__msan_retval_tls
 ; CHECK: ret [2 x i32]
@@ -34,8 +34,8 @@
 
 ; CHECK-LABEL: @InsertValueDouble(
 ; CHECK-DAG: [[Sx:%.*]] = load i64, i64* getelementptr {{.*}}@__msan_param_tls, i32 0, i32 0
-; CHECK: [[A:%.*]] = insertvalue [2 x i64] [i64 -1, i64 -1], i64 [[Sx]], 0
 ; CHECK-DAG: [[Sy:%.*]] = load i64, i64* {{.*}}@__msan_param_tls to i64), i64 8) to i64*)
+; CHECK: [[A:%.*]] = insertvalue [2 x i64] [i64 -1, i64 -1], i64 [[Sx]], 0
 ; CHECK: [[B:%.*]] = insertvalue [2 x i64] [[A]], i64 [[Sy]], 1
 ; CHECK: store [2 x i64] [[B]], [2 x i64]* {{.*}}@__msan_retval_tls
 ; CHECK: ret [2 x double]
diff --git a/llvm/test/Instrumentation/MemorySanitizer/attributes.ll b/llvm/test/Instrumentation/MemorySanitizer/attributes.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/attributes.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/attributes.ll
@@ -49,3 +49,6 @@
 ; CHECK-NOT: writeonly
 ; CHECK-NOT: argmemonly
 ; CHECK-NOT: speculatable
+
+; CHECK: Function Attrs: nounwind readnone willreturn
+; CHECK-NEXT: declare void @llvm.donothing
diff --git a/llvm/test/Instrumentation/MemorySanitizer/clmul.ll b/llvm/test/Instrumentation/MemorySanitizer/clmul.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/clmul.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/clmul.ll
@@ -19,8 +19,8 @@
 
 ; CHECK-LABEL: @clmul00
 ; CHECK: %[[S0:.*]] = load <2 x i64>, <2 x i64>* {{.*}}@__msan_param_tls
-; CHECK: %[[SHUF0:.*]] = shufflevector <2 x i64> %[[S0]], <2 x i64> undef, <2 x i32> zeroinitializer
 ; CHECK: %[[S1:.*]] = load <2 x i64>, <2 x i64>* {{.*}}@__msan_param_tls
+; CHECK: %[[SHUF0:.*]] = shufflevector <2 x i64> %[[S0]], <2 x i64> undef, <2 x i32> zeroinitializer
 ; CHECK: %[[SHUF1:.*]] = shufflevector <2 x i64> %[[S1]], <2 x i64> undef, <2 x i32> zeroinitializer
 ; CHECK: %[[SRET:.*]] = or <2 x i64> %[[SHUF0]], %[[SHUF1]]
 ; CHECK: store <2 x i64> %[[SRET]], <2 x i64>* {{.*}}@__msan_retval_tls
@@ -33,8 +33,8 @@
 
 ; CHECK-LABEL: @clmul10
 ; CHECK: %[[S0:.*]] = load <2 x i64>, <2 x i64>* {{.*}}@__msan_param_tls
-; CHECK: %[[SHUF0:.*]] = shufflevector <2 x i64> %[[S0]], <2 x i64> undef, <2 x i32> zeroinitializer
 ; CHECK: %[[S1:.*]] = load <2 x i64>, <2 x i64>* {{.*}}@__msan_param_tls
+; CHECK: %[[SHUF0:.*]] = shufflevector <2 x i64> %[[S0]], <2 x i64> undef, <2 x i32> zeroinitializer
 ; CHECK: %[[SHUF1:.*]] = shufflevector <2 x i64> %[[S1]], <2 x i64> undef, <2 x i32> <i32 1, i32 1>
 ; CHECK: %[[SRET:.*]] = or <2 x i64> %[[SHUF0]], %[[SHUF1]]
 ; CHECK: store <2 x i64> %[[SRET]], <2 x i64>* {{.*}}@__msan_retval_tls
@@ -47,8 +47,8 @@
 
 ; CHECK-LABEL: @clmul11_256
 ; CHECK: %[[S0:.*]] = load <4 x i64>, <4 x i64>* {{.*}}@__msan_param_tls
-; CHECK: %[[SHUF0:.*]] = shufflevector <4 x i64> %[[S0]], <4 x i64> undef, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
 ; CHECK: %[[S1:.*]] = load <4 x i64>, <4 x i64>* {{.*}}@__msan_param_tls
+; CHECK: %[[SHUF0:.*]] = shufflevector <4 x i64> %[[S0]], <4 x i64> undef, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
 ; CHECK: %[[SHUF1:.*]] = shufflevector <4 x i64> %[[S1]], <4 x i64> undef, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
 ; CHECK: %[[SRET:.*]] = or <4 x i64> %[[SHUF0]], %[[SHUF1]]
 ; CHECK: store <4 x i64> %[[SRET]], <4 x i64>* {{.*}}@__msan_retval_tls
@@ -61,8 +61,8 @@
 
 ; CHECK-LABEL: @clmul01_512
 ; CHECK: %[[S0:.*]] = load <8 x i64>, <8 x i64>* {{.*}}@__msan_param_tls
-; CHECK: %[[SHUF0:.*]] = shufflevector <8 x i64> %[[S0]], <8 x i64> undef, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
 ; CHECK: %[[S1:.*]] = load <8 x i64>, <8 x i64>* {{.*}}@__msan_param_tls
+; CHECK: %[[SHUF0:.*]] = shufflevector <8 x i64> %[[S0]], <8 x i64> undef, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
 ; CHECK: %[[SHUF1:.*]] = shufflevector <8 x i64> %[[S1]], <8 x i64> undef, <8 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7>
 ; CHECK: %[[SRET:.*]] = or <8 x i64> %[[SHUF0]], %[[SHUF1]]
 ; ORIGIN: %[[FLAT:.*]] = bitcast <8 x i64> %[[SHUF1]] to i512
diff --git a/llvm/test/Instrumentation/MemorySanitizer/masked-store-load.ll b/llvm/test/Instrumentation/MemorySanitizer/masked-store-load.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/masked-store-load.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/masked-store-load.ll
@@ -72,14 +72,14 @@
 }
 
 ; CHECK-LABEL: @Load(
+; CHECK: %[[A:.*]] = load <4 x i64>, {{.*}}@__msan_param_tls to i64), i64 8)
+; CHECK-ORIGIN: %[[O:.*]] = load i32, {{.*}}@__msan_param_origin_tls to i64), i64 8)
 ; CHECK: %[[B:.*]] = ptrtoint <4 x double>* %p to i64
 ; CHECK: %[[C:.*]] = xor i64 %[[B]], 87960930222080
 ; CHECK: %[[D:.*]] = inttoptr i64 %[[C]] to <4 x i64>*
 ; CHECK-ORIGIN: %[[E:.*]] = add i64 %[[C]], 17592186044416
 ; CHECK-ORIGIN: %[[F:.*]] = and i64 %[[E]], -4
 ; CHECK-ORIGIN: %[[G:.*]] = inttoptr i64 %[[F]] to i32*
-; CHECK: %[[A:.*]] = load <4 x i64>, {{.*}}@__msan_param_tls to i64), i64 8)
-; CHECK-ORIGIN: %[[O:.*]] = load i32, {{.*}}@__msan_param_origin_tls to i64), i64 8)
 ; CHECK: %[[E:.*]] = call <4 x i64> @llvm.masked.load.v4i64.p0v4i64(<4 x i64>* %[[D]], i32 1, <4 x i1> %mask, <4 x i64> %[[A]])
 ; CHECK-ORIGIN: %[[H:.*]] = load i32, i32* %[[G]]
 ; CHECK-ORIGIN: %[[O2:.*]] = select i1 %{{.*}}, i32 %[[O]], i32 %[[H]]
diff --git a/llvm/test/Instrumentation/MemorySanitizer/msan_basic.ll b/llvm/test/Instrumentation/MemorySanitizer/msan_basic.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/msan_basic.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/msan_basic.ll
@@ -257,6 +257,7 @@
 
 define void @atomic_memcpy(i8* nocapture %x, i8* nocapture %y) nounwind {
   ; CHECK-LABEL: atomic_memcpy
+  ; CHECK-NEXT: call void @llvm.donothing
   ; CHECK-NEXT: call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i64(i8* align 1 %x, i8* align 2 %y, i64 16, i32 1)
   ; CHECK-NEXT: ret void
   call void @llvm.memcpy.element.unordered.atomic.p0i8.p0i8.i64(i8* align 1 %x, i8* align 2 %y, i64 16, i32 1)
@@ -265,6 +266,7 @@
 
 define void @atomic_memmove(i8* nocapture %x, i8* nocapture %y) nounwind {
   ; CHECK-LABEL: atomic_memmove
+  ; CHECK-NEXT: call void @llvm.donothing
   ; CHECK-NEXT: call void @llvm.memmove.element.unordered.atomic.p0i8.p0i8.i64(i8* align 1 %x, i8* align 2 %y, i64 16, i32 1)
   ; CHECK-NEXT: ret void
   call void @llvm.memmove.element.unordered.atomic.p0i8.p0i8.i64(i8* align 1 %x, i8* align 2 %y, i64 16, i32 1)
@@ -273,6 +275,7 @@
 
 define void @atomic_memset(i8* nocapture %x) nounwind {
   ; CHECK-LABEL: atomic_memset
+  ; CHECK-NEXT: call void @llvm.donothing
   ; CHECK-NEXT: call void @llvm.memset.element.unordered.atomic.p0i8.i64(i8* align 1 %x, i8 88, i64 16, i32 1)
   ; CHECK-NEXT: ret void
   call void @llvm.memset.element.unordered.atomic.p0i8.i64(i8* align 1 %x, i8 88, i64 16, i32 1)
@@ -389,6 +392,7 @@
 ; CHECK-LABEL: @IntToPtr
 ; CHECK: load i64, i64*{{.*}}__msan_param_tls
 ; CHECK-ORIGINS-NEXT: load i32, i32*{{.*}}__msan_param_origin_tls
+; CHECK-NEXT: call void @llvm.donothing
 ; CHECK-NEXT: inttoptr
 ; CHECK-NEXT: store i64{{.*}}__msan_retval_tls
 ; CHECK: ret i8*
diff --git a/llvm/test/Instrumentation/MemorySanitizer/msan_eager.ll b/llvm/test/Instrumentation/MemorySanitizer/msan_eager.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/msan_eager.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/msan_eager.ll
@@ -7,6 +7,7 @@
 
 define noundef i32 @NormalRet() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @NormalRet(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    ret i32 123
 ;
   ret i32 123
@@ -14,6 +15,7 @@
 
 define i32 @PartialRet() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @PartialRet(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    store i32 0, i32* bitcast ([100 x i64]* @__msan_retval_tls to i32*), align 8
 ; CHECK-NEXT:    store i32 0, i32* @__msan_retval_origin_tls, align 4
 ; CHECK-NEXT:    ret i32 123
@@ -23,6 +25,7 @@
 
 define noundef i32 @LoadedRet() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @LoadedRet(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    [[P:%.*]] = inttoptr i64 0 to i32*
 ; CHECK-NEXT:    [[O:%.*]] = load i32, i32* [[P]], align 4
 ; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint i32* [[P]] to i64
@@ -35,7 +38,7 @@
 ; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i32 [[_MSLD]], 0
 ; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof !0
 ; CHECK:       7:
-; CHECK-NEXT:    call void @__msan_warning_with_origin_noreturn(i32 [[TMP6]]) #1
+; CHECK-NEXT:    call void @__msan_warning_with_origin_noreturn(i32 [[TMP6]]) [[ATTR2:#.*]]
 ; CHECK-NEXT:    unreachable
 ; CHECK:       8:
 ; CHECK-NEXT:    ret i32 [[O]]
@@ -48,6 +51,7 @@
 
 define void @NormalArg(i32 noundef %a) nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @NormalArg(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    [[P:%.*]] = inttoptr i64 0 to i32*
 ; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint i32* [[P]] to i64
 ; CHECK-NEXT:    [[TMP2:%.*]] = xor i64 [[TMP1]], 87960930222080
@@ -67,6 +71,7 @@
 ; CHECK-LABEL: @PartialArg(
 ; CHECK-NEXT:    [[TMP1:%.*]] = load i32, i32* bitcast ([100 x i64]* @__msan_param_tls to i32*), align 8
 ; CHECK-NEXT:    [[TMP2:%.*]] = load i32, i32* getelementptr inbounds ([200 x i32], [200 x i32]* @__msan_param_origin_tls, i32 0, i32 0), align 4
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    [[P:%.*]] = inttoptr i64 0 to i32*
 ; CHECK-NEXT:    [[TMP3:%.*]] = ptrtoint i32* [[P]] to i64
 ; CHECK-NEXT:    [[TMP4:%.*]] = xor i64 [[TMP3]], 87960930222080
@@ -90,8 +95,9 @@
 
 define void @CallNormal() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @CallNormal(
-; CHECK-NEXT:    [[R:%.*]] = call i32 @NormalRet() #0
-; CHECK-NEXT:    call void @NormalArg(i32 [[R]]) #0
+; CHECK-NEXT:    call void @llvm.donothing()
+; CHECK-NEXT:    [[R:%.*]] = call i32 @NormalRet() [[ATTR0:#.*]]
+; CHECK-NEXT:    call void @NormalArg(i32 [[R]]) [[ATTR0]]
 ; CHECK-NEXT:    ret void
 ;
   %r = call i32 @NormalRet() nounwind uwtable sanitize_memory
@@ -101,6 +107,7 @@
 
 define void @CallWithLoaded() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @CallWithLoaded(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    [[P:%.*]] = inttoptr i64 0 to i32*
 ; CHECK-NEXT:    [[O:%.*]] = load i32, i32* [[P]], align 4
 ; CHECK-NEXT:    [[TMP1:%.*]] = ptrtoint i32* [[P]] to i64
@@ -113,10 +120,10 @@
 ; CHECK-NEXT:    [[_MSCMP:%.*]] = icmp ne i32 [[_MSLD]], 0
 ; CHECK-NEXT:    br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof !0
 ; CHECK:       7:
-; CHECK-NEXT:    call void @__msan_warning_with_origin_noreturn(i32 [[TMP6]]) #1
+; CHECK-NEXT:    call void @__msan_warning_with_origin_noreturn(i32 [[TMP6]]) [[ATTR2]]
 ; CHECK-NEXT:    unreachable
 ; CHECK:       8:
-; CHECK-NEXT:    call void @NormalArg(i32 [[O]]) #0
+; CHECK-NEXT:    call void @NormalArg(i32 [[O]]) [[ATTR0]]
 ; CHECK-NEXT:    ret void
 ;
   %p = inttoptr i64 0 to i32 *
@@ -127,13 +134,14 @@
 
 define void @CallPartial() nounwind uwtable sanitize_memory {
 ; CHECK-LABEL: @CallPartial(
+; CHECK-NEXT:    call void @llvm.donothing()
 ; CHECK-NEXT:    store i32 0, i32* bitcast ([100 x i64]* @__msan_retval_tls to i32*), align 8
-; CHECK-NEXT:    [[R:%.*]] = call i32 @PartialRet() #0
+; CHECK-NEXT:    [[R:%.*]] = call i32 @PartialRet() [[ATTR0]]
 ; CHECK-NEXT:    [[_MSRET:%.*]] = load i32, i32* bitcast ([100 x i64]* @__msan_retval_tls to i32*), align 8
 ; CHECK-NEXT:    [[TMP1:%.*]] = load i32, i32* @__msan_retval_origin_tls, align 4
 ; CHECK-NEXT:    store i32 [[_MSRET]], i32* bitcast ([100 x i64]* @__msan_param_tls to i32*), align 8
 ; CHECK-NEXT:    store i32 [[TMP1]], i32* getelementptr inbounds ([200 x i32], [200 x i32]* @__msan_param_origin_tls, i32 0, i32 0), align 4
-; CHECK-NEXT:    call void @PartialArg(i32 [[R]]) #0
+; CHECK-NEXT:    call void @PartialArg(i32 [[R]]) [[ATTR0]]
 ; CHECK-NEXT:    ret void
 ;
   %r = call i32 @PartialRet() nounwind uwtable sanitize_memory
diff --git a/llvm/test/Instrumentation/MemorySanitizer/msan_kernel_basic.ll b/llvm/test/Instrumentation/MemorySanitizer/msan_kernel_basic.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/msan_kernel_basic.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/msan_kernel_basic.ll
@@ -45,11 +45,11 @@
 ; CHECK: [[BASE:%[0-9]+]] = ptrtoint {{.*}} [[PARAM_SHADOW]]
 ; CHECK: [[SHADOW_PTR:%[a-z0-9_]+]] = inttoptr {{.*}} [[BASE]]
 ; CHECK: [[SHADOW:%[a-z0-9]+]] = load i64, i64* [[SHADOW_PTR]]
+; CHECK: [[BASE2:%[0-9]+]] = ptrtoint {{.*}} [[PARAM_SHADOW]]
 ; Load the shadow of %p and check it
 ; CHECK: icmp ne i64 [[SHADOW]]
 ; CHECK: br i1
 ; CHECK: {{^[0-9]+}}:
-; CHECK: [[BASE2:%[0-9]+]] = ptrtoint {{.*}} [[PARAM_SHADOW]]
 ; CHECK: @__msan_metadata_ptr_for_store_1(i8* %p)
 ; CHECK: store i8
 ; If the new shadow is non-zero, jump to __msan_chain_origin()
diff --git a/llvm/test/Instrumentation/MemorySanitizer/msan_x86intrinsics.ll b/llvm/test/Instrumentation/MemorySanitizer/msan_x86intrinsics.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/msan_x86intrinsics.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/msan_x86intrinsics.ll
@@ -63,6 +63,7 @@
 ; CHECK-ORIGINS: load i32, i32* {{.*}} @__msan_param_origin_tls
 ; CHECK-NEXT: load <8 x i16>, <8 x i16>* {{.*}} @__msan_param_tls
 ; CHECK-ORIGINS: load i32, i32* {{.*}} @__msan_param_origin_tls
+; CHECK-NEXT: call void @llvm.donothing
 ; CHECK-NEXT: = or <8 x i16>
 ; CHECK-ORIGINS: = bitcast <8 x i16> {{.*}} to i128
 ; CHECK-ORIGINS-NEXT: = icmp ne i128 {{.*}}, 0
diff --git a/llvm/test/Instrumentation/MemorySanitizer/unsized_type.ll b/llvm/test/Instrumentation/MemorySanitizer/unsized_type.ll
--- a/llvm/test/Instrumentation/MemorySanitizer/unsized_type.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/unsized_type.ll
@@ -18,6 +18,7 @@
 
 ; CHECK: define void @foo
 ; CHECK-NEXT: entry:
+; CHECK-NEXT: call void @llvm.donothing
 ; CHECK-NEXT: %id = call token @llvm.coro.id
 ; CHECK-NEXT: call i1 @llvm.coro.alloc(token %id)
 ; CHECK-NEXT: ret void